Jump to content

Искусственный интеллект

Страница полузащищена
(Перенаправлено с Искусственный интеллект )

Искусственный интеллект ( ИИ ) в самом широком смысле — это интеллект, демонстрируемый машинами , особенно компьютерными системами . Это область исследований в области информатики , которая разрабатывает и изучает методы и программное обеспечение , которые позволяют машинам воспринимать окружающую среду и использовать обучение и интеллект для принятия мер, которые максимизируют их шансы на достижение определенных целей. [ 1 ] Такие машины можно назвать ИИ.

Некоторые известные приложения ИИ включают передовые поисковые системы в Интернете (например, Google Search ); системы рекомендаций (используются YouTube , Amazon и Netflix ); взаимодействие посредством человеческой речи (например, Google Assistant , Siri и Alexa ); автономные транспортные средства (например, Waymo ); генеративные и творческие инструменты (например, ChatGPT , Apple Intelligence и AI art ); и сверхчеловеческая игра и анализ в стратегических играх (например, шахматах и ​​го ). [ нужна ссылка ] Однако многие приложения ИИ не воспринимаются как ИИ: «Множество передовых ИИ перешло в общие приложения, часто не называемые ИИ, потому что, как только что-то становится достаточно полезным и достаточно распространенным, его больше не называют ИИ ». [ 2 ] [ 3 ]

Алан Тьюринг был первым человеком, проведшим серьезные исследования в области, которую он назвал «машинный интеллект». [ 4 ] Искусственный интеллект был основан как академическая дисциплина в 1956 году. [ 5 ] те, кого сейчас считают отцами-основателями ИИ: Джон Маккарти , Марвин Минкси , Натаниэль Рочестер и Клод Шеннон . [ 6 ] [ 7 ] Эта отрасль прошла через несколько циклов оптимизма. [ 8 ] [ 9 ] за которыми следовали периоды разочарования и потери финансирования, известные как зима AI . [ 10 ] [ 11 ] Финансирование и интерес значительно возросли после 2012 года, когда глубокое обучение превзошло все предыдущие методы искусственного интеллекта. [ 12 ] а после 2017 года с архитектурой-трансформером . [ 13 ] Это привело к буму искусственного интеллекта в начале 2020-х годов, когда компании, университеты и лаборатории, преимущественно базирующиеся в Соединенных Штатах, стали пионерами значительных достижений в области искусственного интеллекта . [ 14 ]

Растущее использование искусственного интеллекта в 21 веке влияет на социальный и экономический сдвиг в сторону большей автоматизации , принятия решений на основе данных и интеграции систем искусственного интеллекта в различные экономические сектора и сферы жизни, влияя на рынки труда , здравоохранение , государственное управление. , промышленность , образование , пропаганда и дезинформация . Это поднимает вопросы о долгосрочных последствиях , этических последствиях и рисках ИИ , что побуждает к дискуссиям о регуляторной политике, призванной обеспечить безопасность и преимущества технологии .

Различные области исследований ИИ сосредоточены вокруг конкретных целей и использования определенных инструментов. Традиционные цели исследований ИИ включают рассуждение , представление знаний , планирование , обучение , обработку естественного языка , восприятие и поддержку робототехники . [ а ] Общий интеллект — способность выполнять любую задачу, которую может выполнить человек, как минимум на равном уровне — входит в число долгосрочных целей этой области. [ 15 ]

Для достижения этих целей исследователи ИИ адаптировали и интегрировали широкий спектр методов, включая поисковую и математическую оптимизацию , формальную логику , искусственные нейронные сети и методы, основанные на статистике , исследовании операций и экономике . [ б ] ИИ также опирается на психологию , лингвистику , философию , нейробиологию и другие области. [ 16 ]

Цели

Общая проблема моделирования (или создания) интеллекта разбита на подзадачи. Они состоят из определенных черт или способностей, которые исследователи ожидают от интеллектуальной системы. Описанные ниже черты получили наибольшее внимание и охватывают сферу исследований ИИ. [ а ]

Рассуждение и решение проблем

Ранние исследователи разработали алгоритмы, имитирующие пошаговые рассуждения, которые люди используют, когда решают головоломки или делают логические выводы . [ 17 ] К концу 1980-х и 1990-м годам были разработаны методы работы с неопределенной или неполной информацией с использованием концепций теории вероятности и экономики . [ 18 ]

Многие из этих алгоритмов недостаточны для решения больших задач рассуждения, потому что они испытывают «комбинаторный взрыв»: они становятся экспоненциально медленнее по мере роста задач. [ 19 ] Даже люди редко используют пошаговые выводы, которые могли смоделировать ранние исследования ИИ. Они решают большинство своих проблем, используя быстрые и интуитивные суждения. [ 20 ] Точные и эффективные рассуждения — нерешенная проблема.

Представление знаний

Онтология представляет знания как набор концепций внутри предметной области и отношений между этими концепциями.

Представление знаний и инженерия знаний [ 21 ] позволить программам искусственного интеллекта разумно отвечать на вопросы и делать выводы о реальных фактах. Формальные представления знаний используются при индексировании и поиске на основе контента. [ 22 ] интерпретация сцены, [ 23 ] поддержка принятия клинических решений, [ 24 ] обнаружение знаний (извлечение «интересных» и практических выводов из больших баз данных ), [ 25 ] и другие области. [ 26 ]

База знаний — это совокупность знаний, представленная в форме, которую может использовать программа. Онтология это набор объектов, отношений, концепций и свойств, используемых в определенной области знаний. [ 27 ] Базы знаний должны представлять такие вещи, как объекты, свойства, категории и отношения между объектами; [ 28 ] ситуации, события, состояния и время; [ 29 ] причины и следствия; [ 30 ] знание о знании (то, что мы знаем о том, что знают другие люди); [ 31 ] рассуждения по умолчанию (вещи, которые люди предполагают, являются истинными до тех пор, пока им не скажут иначе, и останутся истинными, даже когда другие факты меняются); [ 32 ] и многие другие аспекты и области знаний.

Среди наиболее сложных проблем представления знаний — широта знаний здравого смысла (набор атомарных фактов, известных среднестатистическому человеку, огромен); [ 33 ] и субсимволическая форма большинства здравомыслящих знаний (многое из того, что знают люди, не представлено в виде «фактов» или «утверждений», которые они могли бы выразить устно). [ 20 ] Существует также сложность приобретения знаний , проблема получения знаний для приложений ИИ. [ с ]

Планирование и принятие решений

«Агент» — это все, что воспринимает и совершает действия в мире. имеет Рациональный агент цели или предпочтения и предпринимает действия для их достижения. [ д ] [ 36 ] При автоматизированном планировании у агента есть конкретная цель. [ 37 ] При автоматизированном принятии решений у агента есть предпочтения: есть ситуации, в которых он предпочел бы оказаться, а некоторых ситуаций он пытается избежать. Агент, принимающий решения, присваивает каждой ситуации число (называемое « полезностью »), которое измеряет, насколько агент предпочитает ее. Для каждого возможного действия он может рассчитать « ожидаемую полезность »: полезность всех возможных результатов действия, взвешенную по вероятности того, что результат произойдет. Затем он может выбрать действие с максимальной ожидаемой полезностью. [ 38 ]

При классическом планировании агент точно знает, каков будет эффект любого действия. [ 39 ] Однако в большинстве задач реального мира агент может не быть уверен в ситуации, в которой он находится (она «неизвестна» или «ненаблюдаема»), и он может не знать наверняка, что произойдет после каждого возможного действия (это не так). «детерминированный»). Он должен выбрать действие, сделав вероятностное предположение, а затем повторно оценить ситуацию, чтобы увидеть, сработало ли действие. [ 40 ]

В некоторых задачах предпочтения агента могут быть неопределенными, особенно если в решении участвуют другие агенты или люди. Их можно изучить (например, с помощью обратного обучения с подкреплением ), или агент может искать информацию для улучшения своих предпочтений. [ 41 ] Теорию ценности информации можно использовать для взвешивания ценности исследовательских или экспериментальных действий. [ 42 ] Пространство возможных будущих действий и ситуаций обычно непреодолимо велико, поэтому агенты должны предпринимать действия и оценивать ситуации, не зная, каким будет результат.

Марковский процесс принятия решений имеет модель перехода , которая описывает вероятность того, что конкретное действие изменит состояние определенным образом, и функцию вознаграждения , которая определяет полезность каждого состояния и стоимость каждого действия. Политика . связывает решение с каждым возможным состоянием Политика может быть рассчитана (например, путем итерации ), быть эвристической или ее можно изучить. [ 43 ]

Теория игр описывает рациональное поведение нескольких взаимодействующих агентов и используется в программах ИИ, которые принимают решения с участием других агентов. [ 44 ]

Обучение

Машинное обучение — это исследование программ, которые могут автоматически улучшить свою производительность при выполнении определенной задачи. [ 45 ] Он был частью ИИ с самого начала. [ и ]

Существует несколько видов машинного обучения. Обучение без учителя анализирует поток данных, находит закономерности и делает прогнозы без какого-либо другого руководства. [ 48 ] Обучение с учителем требует, чтобы человек сначала помечал входные данные, и существует две основные разновидности: классификация (когда программа должна научиться предсказывать, к какой категории принадлежат входные данные) и регрессия (когда программа должна вывести числовую функцию на основе числового ввода). ). [ 49 ]

При обучении с подкреплением агент вознаграждается за хорошие ответы и наказывается за плохие. Агент учится выбирать ответы, которые классифицируются как «хорошие». [ 50 ] Трансферное обучение – это когда знания, полученные при решении одной проблемы, применяются к новой проблеме. [ 51 ] Глубокое обучение — это тип машинного обучения, при котором входные данные передаются через биологически созданные искусственные нейронные сети для всех этих типов обучения. [ 52 ]

Теория компьютерного обучения может оценивать учащихся по сложности вычислений , по сложности выборки (сколько данных требуется) или по другим понятиям оптимизации . [ 53 ]

Обработка естественного языка

Обработка естественного языка (НЛП) [ 54 ] позволяет программам читать, писать и общаться на человеческих языках, таких как английский . Конкретные проблемы включают распознавание речи , синтез речи , машинный перевод , извлечение информации , поиск информации и ответы на вопросы . [ 55 ]

Ранние работы, основанные на Ноама Хомского , порождающей грамматике и семантических сетях имели трудности с устранением смысловой неоднозначности слов. [ ж ] если только они не ограничены небольшими областями, называемыми « микромирами » (из-за проблемы знания здравого смысла). [ 33 ] ). Маргарет Мастерман считала, что именно смысл, а не грамматика, является ключом к пониманию языков и что тезаурусы основой вычислительной структуры языка должны быть , а не словари.

Современные методы глубокого обучения для НЛП включают встраивание слов (представление слов, обычно в виде векторов, кодирующих их значение), [ 56 ] трансформеры (архитектура глубокого обучения с использованием механизма внимания ), [ 57 ] и другие. [ 58 ] В 2019 году генеративные предварительно обученные языковые модели-трансформеры (или «GPT») начали генерировать связный текст. [ 59 ] [ 60 ] и к 2023 году эти модели смогут получать оценки человеческого уровня на экзаменах на получение адвокатского статуса , SAT , GRE и многих других реальных приложениях. [ 61 ]

Восприятие

Машинное восприятие — это способность использовать данные от датчиков (таких как камеры, микрофоны, беспроводные сигналы, активный лидар , гидролокатор, радар и тактильные датчики ) для определения аспектов мира. Компьютерное зрение — это способность анализировать визуальную информацию. [ 62 ]

Область включает в себя распознавание речи , [ 63 ] классификация изображений , [ 64 ] распознавание лиц , распознавание объектов , [ 65 ] отслеживание объектов , [ 66 ] и роботизированное восприятие . [ 67 ]

Социальный интеллект

Kismet , голова робота, созданная в 1990-х годах; машина, способная распознавать и имитировать эмоции [ 68 ]

Аффективные вычисления — это междисциплинарный зонтик, включающий системы, которые распознают, интерпретируют, обрабатывают или моделируют человеческие чувства, эмоции и настроение . [ 69 ] Например, некоторые виртуальные помощники запрограммированы разговаривать разговорно или даже шутливо подшучивать; это делает их более чувствительными к эмоциональной динамике человеческого взаимодействия или иным образом облегчает взаимодействие человека с компьютером .

Однако это дает наивным пользователям нереальное представление об интеллекте существующих компьютерных агентов. [ 70 ] Умеренные успехи, связанные с аффективными вычислениями, включают текстовый анализ настроений и, в последнее время, мультимодальный анализ настроений , в котором ИИ классифицирует аффекты, отображаемые субъектом, записанным на видео. [ 71 ]

Общий интеллект

Машина с общим искусственным интеллектом должна быть способна решать широкий спектр задач с широтой и универсальностью, аналогичной человеческому интеллекту . [ 15 ]

Техники

Исследования ИИ используют широкий спектр методов для достижения вышеуказанных целей. [ б ]

Поиск и оптимизация

ИИ может решить многие проблемы, разумно перебирая множество возможных решений. [ 72 ] В ИИ используются два совершенно разных типа поиска: поиск в пространстве состояний и локальный поиск .

Поиск в пространстве состояний просматривает дерево возможных состояний, чтобы попытаться найти целевое состояние. [ 73 ] Например, алгоритмы планирования просматривают деревья целей и подцелей, пытаясь найти путь к целевой цели. Этот процесс называется анализом средств и результатов . [ 74 ]

Простой исчерпывающий поиск [ 75 ] редко бывают достаточными для большинства реальных задач: пространство поиска (количество мест для поиска) быстро вырастает до астрономических цифр . В результате поиск выполняется слишком медленно или никогда не завершается. [ 19 ] « Эвристика » или «эмпирические правила» могут помочь расставить приоритеты в выборе, который с большей вероятностью приведет к достижению цели. [ 76 ]

Adversarial search is used for game-playing programs, such as chess or Go. It searches through a tree of possible moves and counter-moves, looking for a winning position.[77]

Illustration of gradient descent for 3 different starting points; two parameters (represented by the plan coordinates) are adjusted in order to minimize the loss function (the height)

Local search uses mathematical optimization to find a solution to a problem. It begins with some form of guess and refines it incrementally.[78]

Gradient descent is a type of local search that optimizes a set of numerical parameters by incrementally adjusting them to minimize a loss function. Variants of gradient descent are commonly used to train neural networks.[79]

Another type of local search is evolutionary computation, which aims to iteratively improve a set of candidate solutions by "mutating" and "recombining" them, selecting only the fittest to survive each generation.[80]

Distributed search processes can coordinate via swarm intelligence algorithms. Two popular swarm algorithms used in search are particle swarm optimization (inspired by bird flocking) and ant colony optimization (inspired by ant trails).[81]

Logic

Formal logic is used for reasoning and knowledge representation.[82] Formal logic comes in two main forms: propositional logic (which operates on statements that are true or false and uses logical connectives such as "and", "or", "not" and "implies")[83] and predicate logic (which also operates on objects, predicates and relations and uses quantifiers such as "Every X is a Y" and "There are some Xs that are Ys").[84]

Deductive reasoning in logic is the process of proving a new statement (conclusion) from other statements that are given and assumed to be true (the premises).[85] Proofs can be structured as proof trees, in which nodes are labelled by sentences, and children nodes are connected to parent nodes by inference rules.

Given a problem and a set of premises, problem-solving reduces to searching for a proof tree whose root node is labelled by a solution of the problem and whose leaf nodes are labelled by premises or axioms. In the case of Horn clauses, problem-solving search can be performed by reasoning forwards from the premises or backwards from the problem.[86] In the more general case of the clausal form of first-order logic, resolution is a single, axiom-free rule of inference, in which a problem is solved by proving a contradiction from premises that include the negation of the problem to be solved.[87]

Inference in both Horn clause logic and first-order logic is undecidable, and therefore intractable. However, backward reasoning with Horn clauses, which underpins computation in the logic programming language Prolog, is Turing complete. Moreover, its efficiency is competitive with computation in other symbolic programming languages.[88]

Fuzzy logic assigns a "degree of truth" between 0 and 1. It can therefore handle propositions that are vague and partially true.[89]

Non-monotonic logics, including logic programming with negation as failure, are designed to handle default reasoning.[32] Other specialized versions of logic have been developed to describe many complex domains.

Probabilistic methods for uncertain reasoning

A simple Bayesian network, with the associated conditional probability tables

Many problems in AI (including in reasoning, planning, learning, perception, and robotics) require the agent to operate with incomplete or uncertain information. AI researchers have devised a number of tools to solve these problems using methods from probability theory and economics.[90] Precise mathematical tools have been developed that analyze how an agent can make choices and plan, using decision theory, decision analysis,[91] and information value theory.[92] These tools include models such as Markov decision processes,[93] dynamic decision networks,[94] game theory and mechanism design.[95]

Bayesian networks[96] are a tool that can be used for reasoning (using the Bayesian inference algorithm),[g][98] learning (using the expectation–maximization algorithm),[h][100] planning (using decision networks)[101] and perception (using dynamic Bayesian networks).[94]

Probabilistic algorithms can also be used for filtering, prediction, smoothing, and finding explanations for streams of data, thus helping perception systems analyze processes that occur over time (e.g., hidden Markov models or Kalman filters).[94]

Expectation–maximization clustering of Old Faithful eruption data starts from a random guess but then successfully converges on an accurate clustering of the two physically distinct modes of eruption

Classifiers and statistical learning methods

The simplest AI applications can be divided into two types: classifiers (e.g., "if shiny then diamond"), on one hand, and controllers (e.g., "if diamond then pick up"), on the other hand. Classifiers[102] are functions that use pattern matching to determine the closest match. They can be fine-tuned based on chosen examples using supervised learning. Each pattern (also called an "observation") is labeled with a certain predefined class. All the observations combined with their class labels are known as a data set. When a new observation is received, that observation is classified based on previous experience.[49]

There are many kinds of classifiers in use. The decision tree is the simplest and most widely used symbolic machine learning algorithm.[103] K-nearest neighbor algorithm was the most widely used analogical AI until the mid-1990s, and Kernel methods such as the support vector machine (SVM) displaced k-nearest neighbor in the 1990s.[104] The naive Bayes classifier is reportedly the "most widely used learner"[105] at Google, due in part to its scalability.[106] Neural networks are also used as classifiers.[107]

Artificial neural networks

A neural network is an interconnected group of nodes, akin to the vast network of neurons in the human brain

An artificial neural network is based on a collection of nodes also known as artificial neurons, which loosely model the neurons in a biological brain. It is trained to recognise patterns; once trained, it can recognise those patterns in fresh data. There is an input, at least one hidden layer of nodes and an output. Each node applies a function and once the weight crosses its specified threshold, the data is transmitted to the next layer. A network is typically called a deep neural network if it has at least 2 hidden layers.[107]

Learning algorithms for neural networks use local search to choose the weights that will get the right output for each input during training. The most common training technique is the backpropagation algorithm.[108] Neural networks learn to model complex relationships between inputs and outputs and find patterns in data. In theory, a neural network can learn any function.[109]

In feedforward neural networks the signal passes in only one direction.[110] Recurrent neural networks feed the output signal back into the input, which allows short-term memories of previous input events. Long short term memory is the most successful network architecture for recurrent networks.[111] Perceptrons[112] use only a single layer of neurons, deep learning[113] uses multiple layers. Convolutional neural networks strengthen the connection between neurons that are "close" to each other—this is especially important in image processing, where a local set of neurons must identify an "edge" before the network can identify an object.[114]

Deep learning

Deep learning[113] uses several layers of neurons between the network's inputs and outputs. The multiple layers can progressively extract higher-level features from the raw input. For example, in image processing, lower layers may identify edges, while higher layers may identify the concepts relevant to a human such as digits, letters, or faces.[115]

Deep learning has profoundly improved the performance of programs in many important subfields of artificial intelligence, including computer vision, speech recognition, natural language processing, image classification,[116] and others. The reason that deep learning performs so well in so many applications is not known as of 2023.[117] The sudden success of deep learning in 2012–2015 did not occur because of some new discovery or theoretical breakthrough (deep neural networks and backpropagation had been described by many people, as far back as the 1950s)[i] but because of two factors: the incredible increase in computer power (including the hundred-fold increase in speed by switching to GPUs) and the availability of vast amounts of training data, especially the giant curated datasets used for benchmark testing, such as ImageNet.[j]

GPT

Generative pre-trained transformers (GPT) are large language models (LLMs) that are based on the semantic relationships between words in sentences (natural language processing). Text-based GPT models are pretrained on a large corpus of text that can be from the Internet. The pretraining consists of predicting the next token (a token being usually a word, subword, or punctuation). Throughout this pretraining, GPT models accumulate knowledge about the world and can then generate human-like text by repeatedly predicting the next token. Typically, a subsequent training phase makes the model more truthful, useful, and harmless, usually with a technique called reinforcement learning from human feedback (RLHF). Current GPT models are prone to generating falsehoods called "hallucinations", although this can be reduced with RLHF and quality data. They are used in chatbots, which allow people to ask a question or request a task in simple text.[125][126]

Current models and services include Gemini (formerly Bard), ChatGPT, Grok, Claude, Copilot, and LLaMA.[127] Multimodal GPT models can process different types of data (modalities) such as images, videos, sound, and text.[128]

Specialized hardware and software

In the late 2010s, graphics processing units (GPUs) that were increasingly designed with AI-specific enhancements and used with specialized TensorFlow software had replaced previously used central processing unit (CPUs) as the dominant means for large-scale (commercial and academic) machine learning models' training.[129] Specialized programming languages such as Prolog were used in early AI research,[130] but general-purpose programming languages like Python have become predominant.[131]

Applications

AI and machine learning technology is used in most of the essential applications of the 2020s, including: search engines (such as Google Search), targeting online advertisements, recommendation systems (offered by Netflix, YouTube or Amazon), driving internet traffic, targeted advertising (AdSense, Facebook), virtual assistants (such as Siri or Alexa), autonomous vehicles (including drones, ADAS and self-driving cars), automatic language translation (Microsoft Translator, Google Translate), facial recognition (Apple's Face ID or Microsoft's DeepFace and Google's FaceNet) and image labeling (used by Facebook, Apple's iPhoto and TikTok). The deployment of AI may be overseen by a Chief automation officer (CAO).

Health and medicine

The application of AI in medicine and medical research has the potential to increase patient care and quality of life.[132] Through the lens of the Hippocratic Oath, medical professionals are ethically compelled to use AI, if applications can more accurately diagnose and treat patients.

For medical research, AI is an important tool for processing and integrating big data. This is particularly important for organoid and tissue engineering development which use microscopy imaging as a key technique in fabrication.[133] It has been suggested that AI can overcome discrepancies in funding allocated to different fields of research.[133] New AI tools can deepen the understanding of biomedically relevant pathways. For example, AlphaFold 2 (2021) demonstrated the ability to approximate, in hours rather than months, the 3D structure of a protein.[134] In 2023, it was reported that AI-guided drug discovery helped find a class of antibiotics capable of killing two different types of drug-resistant bacteria.[135] In 2024, researchers used machine learning to accelerate the search for Parkinson's disease drug treatments. Their aim was to identify compounds that block the clumping, or aggregation, of alpha-synuclein (the protein that characterises Parkinson's disease). They were able to speed up the initial screening process ten-fold and reduce the cost by a thousand-fold.[136][137]

Games

Game playing programs have been used since the 1950s to demonstrate and test AI's most advanced techniques.[138] Deep Blue became the first computer chess-playing system to beat a reigning world chess champion, Garry Kasparov, on 11 May 1997.[139] In 2011, in a Jeopardy! quiz show exhibition match, IBM's question answering system, Watson, defeated the two greatest Jeopardy! champions, Brad Rutter and Ken Jennings, by a significant margin.[140] In March 2016, AlphaGo won 4 out of 5 games of Go in a match with Go champion Lee Sedol, becoming the first computer Go-playing system to beat a professional Go player without handicaps. Then in 2017 it defeated Ke Jie, who was the best Go player in the world.[141] Other programs handle imperfect-information games, such as the poker-playing program Pluribus.[142] DeepMind developed increasingly generalistic reinforcement learning models, such as with MuZero, which could be trained to play chess, Go, or Atari games.[143] In 2019, DeepMind's AlphaStar achieved grandmaster level in StarCraft II, a particularly challenging real-time strategy game that involves incomplete knowledge of what happens on the map.[144] In 2021, an AI agent competed in a PlayStation Gran Turismo competition, winning against four of the world's best Gran Turismo drivers using deep reinforcement learning.[145] In 2024, Google DeepMind introduced SIMA, a type of AI capable of autonomously playing nine previously unseen open-world video games by observing screen output, as well as executing short, specific tasks in response to natural language instructions.[146]

Mathematics

In mathematics, special forms of formal step-by-step reasoning are used. In contrast, LLMs such as GPT-4 Turbo, Gemini Ultra, Claude Opus, LLaMa-2 or Mistral Large are working with probabilistic models, which can produce wrong answers in the form of hallucinations. Therefore, they need not only a large database of mathematical problems to learn from but also methods such as supervised fine-tuning or trained classifiers with human-annotated data to improve answers for new problems and learn from corrections.[147]

Alternatively, dedicated models for mathematic problem solving with higher precision for the outcome including proof of theorems have been developed such as Alpha Tensor, Alpha Geometry and Alpha Proof all from Google DeepMind,[148] Llemma from eleuther[149] or Julius.[150]

When natural language is used to describe mathematical problems, converters transform such prompts into a formal language such as Lean to define mathematic tasks.

Some models have been developed to solve challenging problems and reach good results in benchmark tests, others to serve as educational tools in mathematics.[151]

Finance

Finance is one of the fastest growing sectors where applied AI tools are being deployed: from retail online banking to investment advice and insurance, where automated "robot advisers" have been in use for some years. [152]

World Pensions experts like Nicolas Firzli insist it may be too early to see the emergence of highly innovative AI-informed financial products and services: "the deployment of AI tools will simply further automatise things: destroying tens of thousands of jobs in banking, financial planning, and pension advice in the process, but I’m not sure it will unleash a new wave of [e.g., sophisticated] pension innovation."[153]

Military

Various countries are deploying AI military applications.[154] The main applications enhance command and control, communications, sensors, integration and interoperability.[155] Research is targeting intelligence collection and analysis, logistics, cyber operations, information operations, and semiautonomous and autonomous vehicles.[154] AI technologies enable coordination of sensors and effectors, threat detection and identification, marking of enemy positions, target acquisition, coordination and deconfliction of distributed Joint Fires between networked combat vehicles involving manned and unmanned teams.[155] AI was incorporated into military operations in Iraq and Syria.[154]

In November 2023, US Vice President Kamala Harris disclosed a declaration signed by 31 nations to set guardrails for the military use of AI. The commitments include using legal reviews to ensure the compliance of military AI with international laws, and being cautious and transparent in the development of this technology.[156]

Generative AI

Vincent van Gogh in watercolour created by generative AI software

In the early 2020s, generative AI gained widespread prominence. In March 2023, 58% of U.S. adults had heard about ChatGPT and 14% had tried it.[157] The increasing realism and ease-of-use of AI-based text-to-image generators such as Midjourney, DALL-E, and Stable Diffusion sparked a trend of viral AI-generated photos. Widespread attention was gained by a fake photo of Pope Francis wearing a white puffer coat, the fictional arrest of Donald Trump, and a hoax of an attack on the Pentagon, as well as the usage in professional creative arts.[158][159]

Other industry-specific tasks

There are also thousands of successful AI applications used to solve specific problems for specific industries or institutions. In a 2017 survey, one in five companies reported having incorporated "AI" in some offerings or processes.[160] A few examples are energy storage, medical diagnosis, military logistics, applications that predict the result of judicial decisions, foreign policy, or supply chain management.

AI applications for evacuation and disaster management are growing. AI has been used to investigate if and how people evacuated in large scale and small scale evacuations using historical data from GPS, videos or social media. Further, AI can provide real time information on the real time evacuation conditions.[161][162][163]

In agriculture, AI has helped farmers identify areas that need irrigation, fertilization, pesticide treatments or increasing yield. Agronomists use AI to conduct research and development. AI has been used to predict the ripening time for crops such as tomatoes, monitor soil moisture, operate agricultural robots, conduct predictive analytics, classify livestock pig call emotions, automate greenhouses, detect diseases and pests, and save water.

Artificial intelligence is used in astronomy to analyze increasing amounts of available data and applications, mainly for "classification, regression, clustering, forecasting, generation, discovery, and the development of new scientific insights" for example for discovering exoplanets, forecasting solar activity, and distinguishing between signals and instrumental effects in gravitational wave astronomy. It could also be used for activities in space such as space exploration, including analysis of data from space missions, real-time science decisions of spacecraft, space debris avoidance, and more autonomous operation.

Ethics

AI has potential benefits and potential risks. AI may be able to advance science and find solutions for serious problems: Demis Hassabis of Deep Mind hopes to "solve intelligence, and then use that to solve everything else".[164] However, as the use of AI has become widespread, several unintended consequences and risks have been identified.[165] In-production systems can sometimes not factor ethics and bias into their AI training processes, especially when the AI algorithms are inherently unexplainable in deep learning.[166]

Risks and harm

Machine-learning algorithms require large amounts of data. The techniques used to acquire this data have raised concerns about privacy, surveillance and copyright.

Technology companies collect a wide range of data from their users, including online activity, geolocation data, video and audio.[167] For example, in order to build speech recognition algorithms, Amazon has recorded millions of private conversations and allowed temporary workers to listen to and transcribe some of them.[168] Opinions about this widespread surveillance range from those who see it as a necessary evil to those for whom it is clearly unethical and a violation of the right to privacy.[169]

AI developers argue that this is the only way to deliver valuable applications. and have developed several techniques that attempt to preserve privacy while still obtaining the data, such as data aggregation, de-identification and differential privacy.[170] Since 2016, some privacy experts, such as Cynthia Dwork, have begun to view privacy in terms of fairness. Brian Christian wrote that experts have pivoted "from the question of 'what they know' to the question of 'what they're doing with it'."[171]

Generative AI is often trained on unlicensed copyrighted works, including in domains such as images or computer code; the output is then used under the rationale of "fair use". Experts disagree about how well and under what circumstances this rationale will hold up in courts of law; relevant factors may include "the purpose and character of the use of the copyrighted work" and "the effect upon the potential market for the copyrighted work".[172][173] Website owners who do not wish to have their content scraped can indicate it in a "robots.txt" file.[174] In 2023, leading authors (including John Grisham and Jonathan Franzen) sued AI companies for using their work to train generative AI.[175][176] Another discussed approach is to envision a separate sui generis system of protection for creations generated by AI to ensure fair attribution and compensation for human authors.[177]

Dominance by tech giants

The commercial AI scene is dominated by Big Tech companies such as Alphabet Inc., Amazon, Apple Inc., Meta Platforms, and Microsoft.[178][179][180] Some of these players already own the vast majority of existing cloud infrastructure and computing power from data centers, allowing them to entrench further in the marketplace.[181][182]

Substantial power needs and other environmental impacts

In January 2024, the International Energy Agency (IEA) released Electricity 2024, Analysis and Forecast to 2026, forecasting electric power use.[183] This is the first IEA report to make projections for data centers and power consumption for artificial intelligence and cryptocurrency. The report states that power demand for these uses might double by 2026, with additional electric power usage equal to electricity used by the whole Japanese nation.[184]

Prodigious power consumption by AI is responsible for the growth of fossil fuels use, and might delay closings of obsolete, carbon-emitting coal energy facilities. There is a feverish rise in the construction of data centers throughout the US, making large technology firms (e.g., Microsoft, Meta, Google, Amazon) into voracious consumers of electric power. Projected electric consumption is so immense that there is concern that it will be fulfilled no matter the source. A ChatGPT search involves the use of 10 times the electrical energy as a Google search. The large firms are in haste to find power sources – from nuclear energy to geothermal to fusion. The tech firms argue that – in the long view – AI will be eventually kinder to the environment, but they need the energy now. AI makes the power grid more efficient and "intelligent", will assist in the growth of nuclear power, and track overall carbon emissions, according to technology firms.[185]

A 2024 Goldman Sachs Research Paper, AI Data Centers and the Coming US Power Demand Surge, found "US power demand (is) likely to experience growth not seen in a generation…." and forecasts that, by 2030, US data centers will consume 8% of US power, as opposed to 3% in 2022, presaging growth for the electrical power generation industry by a variety of means.[186]Data centers' need for more and more electrical power is such that they might max out the electrical grid. The Big Tech companies counter that AI can be used to maximize the utilization of the grid by all.[187]

In 2024, the Wall Street Journal reported that big AI companies have begun negotiations with the US nuclear power providers to provide electricity to the data centers. In March 2024 Amazon purchased a Pennsylvania nuclear-powered data center for $650 Million (US).[188]

Misinformation

YouTube, Facebook and others use recommender systems to guide users to more content. These AI programs were given the goal of maximizing user engagement (that is, the only goal was to keep people watching). The AI learned that users tended to choose misinformation, conspiracy theories, and extreme partisan content, and, to keep them watching, the AI recommended more of it. Users also tended to watch more content on the same subject, so the AI led people into filter bubbles where they received multiple versions of the same misinformation.[189] This convinced many users that the misinformation was true, and ultimately undermined trust in institutions, the media and the government.[190] The AI program had correctly learned to maximize its goal, but the result was harmful to society. After the U.S. election in 2016, major technology companies took steps to mitigate the problem [citation needed].

In 2022, generative AI began to create images, audio, video and text that are indistinguishable from real photographs, recordings, films, or human writing. It is possible for bad actors to use this technology to create massive amounts of misinformation or propaganda.[191] AI pioneer Geoffrey Hinton expressed concern about AI enabling "authoritarian leaders to manipulate their electorates" on a large scale, among other risks.[192]

Algorithmic bias and fairness

In statistics, a bias is a systematic error or deviation from the correct value. But in the context of fairness, it often refers to a tendency in favor or against a certain group or individual characteristic, usually in a way that is considered unfair or harmful. A statistically unbiased AI system that produces disparate outcomes for different demographic groups may thus be viewed as biased in the ethical sense.[193]

The field of fairness studies how to prevent harms from algorithmic biases. There are various conflicting definitions and mathematical models of fairness. These notions depend on ethical assumptions, and are influenced by beliefs about society. One broad category is distributive fairness, which focuses on the outcomes, often identifying groups and seeking to compensate for statistical disparities. Representational fairness tries to ensure that AI systems don't reinforce negative stereotypes or render certain groups invisible. Procedural fairness focuses on the decision process rather than the outcome. The most relevant notions of fairness may depend on the context, notably the type of AI application and the stakeholders. The subjectivity in the notions of bias and fairness makes it difficult for companies to operationalize them. Having access to sensitive attributes such as race or gender is also considered by many AI ethicists to be necessary in order to compensate for biases, but it may conflict with anti-discrimination laws.[193]

Machine learning applications will be biased if they learn from biased data.[194] The developers may not be aware that the bias exists.[195] Bias can be introduced by the way training data is selected and by the way a model is deployed.[196][194] If a biased algorithm is used to make decisions that can seriously harm people (as it can in medicine, finance, recruitment, housing or policing) then the algorithm may cause discrimination.[197]

On June 28, 2015, Google Photos's new image labeling feature mistakenly identified Jacky Alcine and a friend as "gorillas" because they were black. The system was trained on a dataset that contained very few images of black people,[198] a problem called "sample size disparity".[199] Google "fixed" this problem by preventing the system from labelling anything as a "gorilla". Eight years later, in 2023, Google Photos still could not identify a gorilla, and neither could similar products from Apple, Facebook, Microsoft and Amazon.[200]

COMPAS is a commercial program widely used by U.S. courts to assess the likelihood of a defendant becoming a recidivist. In 2016, Julia Angwin at ProPublica discovered that COMPAS exhibited racial bias, despite the fact that the program was not told the races of the defendants. Although the error rate for both whites and blacks was calibrated equal at exactly 61%, the errors for each race were different—the system consistently overestimated the chance that a black person would re-offend and would underestimate the chance that a white person would not re-offend.[201] In 2017, several researchers[k] showed that it was mathematically impossible for COMPAS to accommodate all possible measures of fairness when the base rates of re-offense were different for whites and blacks in the data.[203]

A program can make biased decisions even if the data does not explicitly mention a problematic feature (such as "race" or "gender"). The feature will correlate with other features (like "address", "shopping history" or "first name"), and the program will make the same decisions based on these features as it would on "race" or "gender".[204] Moritz Hardt said "the most robust fact in this research area is that fairness through blindness doesn't work."[205]

Criticism of COMPAS highlighted that machine learning models are designed to make "predictions" that are only valid if we assume that the future will resemble the past. If they are trained on data that includes the results of racist decisions in the past, machine learning models must predict that racist decisions will be made in the future. If an application then uses these predictions as recommendations, some of these "recommendations" will likely be racist.[206] Thus, machine learning is not well suited to help make decisions in areas where there is hope that the future will be better than the past. It is descriptive rather than prescriptive.[l]

Bias and unfairness may go undetected because the developers are overwhelmingly white and male: among AI engineers, about 4% are black and 20% are women.[199]

At its 2022 Conference on Fairness, Accountability, and Transparency (ACM FAccT 2022), the Association for Computing Machinery, in Seoul, South Korea, presented and published findings that recommend that until AI and robotics systems are demonstrated to be free of bias mistakes, they are unsafe, and the use of self-learning neural networks trained on vast, unregulated sources of flawed internet data should be curtailed.[dubiousdiscuss][208]

Lack of transparency

Many AI systems are so complex that their designers cannot explain how they reach their decisions.[209] Particularly with deep neural networks, in which there are a large amount of non-linear relationships between inputs and outputs. But some popular explainability techniques exist.[210]

It is impossible to be certain that a program is operating correctly if no one knows how exactly it works. There have been many cases where a machine learning program passed rigorous tests, but nevertheless learned something different than what the programmers intended. For example, a system that could identify skin diseases better than medical professionals was found to actually have a strong tendency to classify images with a ruler as "cancerous", because pictures of malignancies typically include a ruler to show the scale.[211] Another machine learning system designed to help effectively allocate medical resources was found to classify patients with asthma as being at "low risk" of dying from pneumonia. Having asthma is actually a severe risk factor, but since the patients having asthma would usually get much more medical care, they were relatively unlikely to die according to the training data. The correlation between asthma and low risk of dying from pneumonia was real, but misleading.[212]

People who have been harmed by an algorithm's decision have a right to an explanation.[213] Doctors, for example, are expected to clearly and completely explain to their colleagues the reasoning behind any decision they make. Early drafts of the European Union's General Data Protection Regulation in 2016 included an explicit statement that this right exists.[m] Industry experts noted that this is an unsolved problem with no solution in sight. Regulators argued that nevertheless the harm is real: if the problem has no solution, the tools should not be used.[214]

DARPA established the XAI ("Explainable Artificial Intelligence") program in 2014 to try and solve these problems.[215]

Several approaches aim to address the transparency problem. SHAP enables to visualise the contribution of each feature to the output.[216] LIME can locally approximate a model's outputs with a simpler, interpretable model.[217] Multitask learning provides a large number of outputs in addition to the target classification. These other outputs can help developers deduce what the network has learned.[218] Deconvolution, DeepDream and other generative methods can allow developers to see what different layers of a deep network for computer vision have learned, and produce output that can suggest what the network is learning.[219] For generative pre-trained transformers, Anthropic developed a technique based on dictionary learning that associates patterns of neuron activations with human-understandable concepts.[220]

Bad actors and weaponized AI

Artificial intelligence provides a number of tools that are useful to bad actors, such as authoritarian governments, terrorists, criminals or rogue states.

A lethal autonomous weapon is a machine that locates, selects and engages human targets without human supervision.[n] Widely available AI tools can be used by bad actors to develop inexpensive autonomous weapons and, if produced at scale, they are potentially weapons of mass destruction.[222] Even when used in conventional warfare, it is unlikely that they will be unable to reliably choose targets and could potentially kill an innocent person.[222] In 2014, 30 nations (including China) supported a ban on autonomous weapons under the United Nations' Convention on Certain Conventional Weapons, however the United States and others disagreed.[223] By 2015, over fifty countries were reported to be researching battlefield robots.[224]

AI tools make it easier for authoritarian governments to efficiently control their citizens in several ways. Face and voice recognition allow widespread surveillance. Machine learning, operating this data, can classify potential enemies of the state and prevent them from hiding. Recommendation systems can precisely target propaganda and misinformation for maximum effect. Deepfakes and generative AI aid in producing misinformation. Advanced AI can make authoritarian centralized decision making more competitive than liberal and decentralized systems such as markets. It lowers the cost and difficulty of digital warfare and advanced spyware.[225] All these technologies have been available since 2020 or earlier—AI facial recognition systems are already being used for mass surveillance in China.[226][227]

There many other ways that AI is expected to help bad actors, some of which can not be foreseen. For example, machine-learning AI is able to design tens of thousands of toxic molecules in a matter of hours.[228]

Technological unemployment

Economists have frequently highlighted the risks of redundancies from AI, and speculated about unemployment if there is no adequate social policy for full employment.[229]

In the past, technology has tended to increase rather than reduce total employment, but economists acknowledge that "we're in uncharted territory" with AI.[230] A survey of economists showed disagreement about whether the increasing use of robots and AI will cause a substantial increase in long-term unemployment, but they generally agree that it could be a net benefit if productivity gains are redistributed.[231] Risk estimates vary; for example, in the 2010s, Michael Osborne and Carl Benedikt Frey estimated 47% of U.S. jobs are at "high risk" of potential automation, while an OECD report classified only 9% of U.S. jobs as "high risk".[o][233] The methodology of speculating about future employment levels has been criticised as lacking evidential foundation, and for implying that technology, rather than social policy, creates unemployment, as opposed to redundancies.[229] In April 2023, it was reported that 70% of the jobs for Chinese video game illustrators had been eliminated by generative artificial intelligence.[234][235]

Unlike previous waves of automation, many middle-class jobs may be eliminated by artificial intelligence; The Economist stated in 2015 that "the worry that AI could do to white-collar jobs what steam power did to blue-collar ones during the Industrial Revolution" is "worth taking seriously".[236] Jobs at extreme risk range from paralegals to fast food cooks, while job demand is likely to increase for care-related professions ranging from personal healthcare to the clergy.[237]

From the early days of the development of artificial intelligence, there have been arguments, for example, those put forward by Joseph Weizenbaum, about whether tasks that can be done by computers actually should be done by them, given the difference between computers and humans, and between quantitative calculation and qualitative, value-based judgement.[238]

Existential risk

It has been argued AI will become so powerful that humanity may irreversibly lose control of it. This could, as physicist Stephen Hawking stated, "spell the end of the human race".[239] This scenario has been common in science fiction, when a computer or robot suddenly develops a human-like "self-awareness" (or "sentience" or "consciousness") and becomes a malevolent character.[p] These sci-fi scenarios are misleading in several ways.

First, AI does not require human-like "sentience" to be an existential risk. Modern AI programs are given specific goals and use learning and intelligence to achieve them. Philosopher Nick Bostrom argued that if one gives almost any goal to a sufficiently powerful AI, it may choose to destroy humanity to achieve it (he used the example of a paperclip factory manager).[241] Stuart Russell gives the example of household robot that tries to find a way to kill its owner to prevent it from being unplugged, reasoning that "you can't fetch the coffee if you're dead."[242] In order to be safe for humanity, a superintelligence would have to be genuinely aligned with humanity's morality and values so that it is "fundamentally on our side".[243]

Second, Yuval Noah Harari argues that AI does not require a robot body or physical control to pose an existential risk. The essential parts of civilization are not physical. Things like ideologies, law, government, money and the economy are made of language; they exist because there are stories that billions of people believe. The current prevalence of misinformation suggests that an AI could use language to convince people to believe anything, even to take actions that are destructive.[244]

The opinions amongst experts and industry insiders are mixed, with sizable fractions both concerned and unconcerned by risk from eventual superintelligent AI.[245] Personalities such as Stephen Hawking, Bill Gates, and Elon Musk,[246] as well as AI pioneers such as Yoshua Bengio, Stuart Russell, Demis Hassabis, and Sam Altman, have expressed concerns about existential risk from AI.

In May 2023, Geoffrey Hinton announced his resignation from Google in order to be able to "freely speak out about the risks of AI" without "considering how this impacts Google."[247] He notably mentioned risks of an AI takeover,[248] and stressed that in order to avoid the worst outcomes, establishing safety guidelines will require cooperation among those competing in use of AI.[249]

In 2023, many leading AI experts issued the joint statement that "Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war".[250]

Other researchers, however, spoke in favor of a less dystopian view. AI pioneer Juergen Schmidhuber did not sign the joint statement, emphasising that in 95% of all cases, AI research is about making "human lives longer and healthier and easier."[251] While the tools that are now being used to improve lives can also be used by bad actors, "they can also be used against the bad actors."[252][253] Andrew Ng also argued that "it's a mistake to fall for the doomsday hype on AI—and that regulators who do will only benefit vested interests."[254] Yann LeCun "scoffs at his peers' dystopian scenarios of supercharged misinformation and even, eventually, human extinction."[255] In the early 2010s, experts argued that the risks are too distant in the future to warrant research or that humans will be valuable from the perspective of a superintelligent machine.[256] However, after 2016, the study of current and future risks and possible solutions became a serious area of research.[257]

Ethical machines and alignment

Friendly AI are machines that have been designed from the beginning to minimize risks and to make choices that benefit humans. Eliezer Yudkowsky, who coined the term, argues that developing friendly AI should be a higher research priority: it may require a large investment and it must be completed before AI becomes an existential risk.[258]

Machines with intelligence have the potential to use their intelligence to make ethical decisions. The field of machine ethics provides machines with ethical principles and procedures for resolving ethical dilemmas.[259] The field of machine ethics is also called computational morality,[259] and was founded at an AAAI symposium in 2005.[260]

Other approaches include Wendell Wallach's "artificial moral agents"[261] and Stuart J. Russell's three principles for developing provably beneficial machines.[262]

Open source

Active organizations in the AI open-source community include Hugging Face,[263] Google,[264] EleutherAI and Meta.[265] Various AI models, such as Llama 2, Mistral or Stable Diffusion, have been made open-weight,[266][267] meaning that their architecture and trained parameters (the "weights") are publicly available. Open-weight models can be freely fine-tuned, which allows companies to specialize them with their own data and for their own use-case.[268] Open-weight models are useful for research and innovation but can also be misused. Since they can be fine-tuned, any built-in security measure, such as objecting to harmful requests, can be trained away until it becomes ineffective. Some researchers warn that future AI models may develop dangerous capabilities (such as the potential to drastically facilitate bioterrorism) and that once released on the Internet, they can't be deleted everywhere if needed. They recommend pre-release audits and cost-benefit analyses.[269]

Frameworks

Artificial Intelligence projects can have their ethical permissibility tested while designing, developing, and implementing an AI system. An AI framework such as the Care and Act Framework containing the SUM values—developed by the Alan Turing Institute tests projects in four main areas:[270][271]

  • Respect the dignity of individual people
  • Connect with other people sincerely, openly, and inclusively
  • Care for the wellbeing of everyone
  • Protect social values, justice, and the public interest

Other developments in ethical frameworks include those decided upon during the Asilomar Conference, the Montreal Declaration for Responsible AI, and the IEEE's Ethics of Autonomous Systems initiative, among others;[272] however, these principles do not go without their criticisms, especially regards to the people chosen contributes to these frameworks.[273]

Promotion of the wellbeing of the people and communities that these technologies affect requires consideration of the social and ethical implications at all stages of AI system design, development and implementation, and collaboration between job roles such as data scientists, product managers, data engineers, domain experts, and delivery managers.[274]

The UK AI Safety Institute released in 2024 a testing toolset called 'Inspect' for AI safety evaluations available under a MIT open-source licence which is freely available on GitHub and can be improved with third-party packages. It can be used to evaluate AI models in a range of areas including core knowledge, ability to reason, and autonomous capabilities.[275]

Regulation

AI Safety Summit
The first global AI Safety Summit was held in 2023 with a declaration calling for international co-operation

The regulation of artificial intelligence is the development of public sector policies and laws for promoting and regulating AI; it is therefore related to the broader regulation of algorithms.[276] The regulatory and policy landscape for AI is an emerging issue in jurisdictions globally.[277] According to AI Index at Stanford, the annual number of AI-related laws passed in the 127 survey countries jumped from one passed in 2016 to 37 passed in 2022 alone.[278][279] Between 2016 and 2020, more than 30 countries adopted dedicated strategies for AI.[280] Most EU member states had released national AI strategies, as had Canada, China, India, Japan, Mauritius, the Russian Federation, Saudi Arabia, United Arab Emirates, U.S., and Vietnam. Others were in the process of elaborating their own AI strategy, including Bangladesh, Malaysia and Tunisia.[280] The Global Partnership on Artificial Intelligence was launched in June 2020, stating a need for AI to be developed in accordance with human rights and democratic values, to ensure public confidence and trust in the technology.[280] Henry Kissinger, Eric Schmidt, and Daniel Huttenlocher published a joint statement in November 2021 calling for a government commission to regulate AI.[281] In 2023, OpenAI leaders published recommendations for the governance of superintelligence, which they believe may happen in less than 10 years.[282] In 2023, the United Nations also launched an advisory body to provide recommendations on AI governance; the body comprises technology company executives, governments officials and academics.[283]

In a 2022 Ipsos survey, attitudes towards AI varied greatly by country; 78% of Chinese citizens, but only 35% of Americans, agreed that "products and services using AI have more benefits than drawbacks".[278] A 2023 Reuters/Ipsos poll found that 61% of Americans agree, and 22% disagree, that AI poses risks to humanity.[284] In a 2023 Fox News poll, 35% of Americans thought it "very important", and an additional 41% thought it "somewhat important", for the federal government to regulate AI, versus 13% responding "not very important" and 8% responding "not at all important".[285][286]

In November 2023, the first global AI Safety Summit was held in Bletchley Park in the UK to discuss the near and far term risks of AI and the possibility of mandatory and voluntary regulatory frameworks.[287] 28 countries including the United States, China, and the European Union issued a declaration at the start of the summit, calling for international co-operation to manage the challenges and risks of artificial intelligence.[288][289] In May 2024 at the AI Seoul Summit, 16 global AI tech companies agreed to safety commitments on the development of AI.[290][291]

History

The study of mechanical or "formal" reasoning began with philosophers and mathematicians in antiquity. The study of logic led directly to Alan Turing's theory of computation, which suggested that a machine, by shuffling symbols as simple as "0" and "1", could simulate any conceivable form of mathematical reasoning.[292][4] This, along with concurrent discoveries in cybernetics, information theory and neurobiology, led researchers to consider the possibility of building an "electronic brain".[q] They developed several areas of research that would become part of AI,[294] such as McCullouch and Pitts design for "artificial neurons" in 1943,[118] and Turing's influential 1950 paper 'Computing Machinery and Intelligence', which introduced the Turing test and showed that "machine intelligence" was plausible.[295][4]

The field of AI research was founded at a workshop at Dartmouth College in 1956.[r][5] The attendees became the leaders of AI research in the 1960s.[s] They and their students produced programs that the press described as "astonishing":[t] computers were learning checkers strategies, solving word problems in algebra, proving logical theorems and speaking English.[u][8] Artificial intelligence laboratories were set up at a number of British and U.S. universities in the latter 1950s and early 1960s.[4]

Researchers in the 1960s and the 1970s were convinced that their methods would eventually succeed in creating a machine with general intelligence and considered this the goal of their field.[299] In 1965 Herbert Simon predicted, "machines will be capable, within twenty years, of doing any work a man can do".[300] In 1967 Marvin Minsky agreed, writing, "within a generation ... the problem of creating 'artificial intelligence' will substantially be solved".[301] They had, however, underestimated the difficulty of the problem.[v] In 1974, both the U.S. and British governments cut off exploratory research in response to the criticism of Sir James Lighthill[303] and ongoing pressure from the U.S. Congress to fund more productive projects.[304] Minsky's and Papert's book Perceptrons was understood as proving that artificial neural networks would never be useful for solving real-world tasks, thus discrediting the approach altogether.[305] The "AI winter", a period when obtaining funding for AI projects was difficult, followed.[10]

In the early 1980s, AI research was revived by the commercial success of expert systems,[306] a form of AI program that simulated the knowledge and analytical skills of human experts. By 1985, the market for AI had reached over a billion dollars. At the same time, Japan's fifth generation computer project inspired the U.S. and British governments to restore funding for academic research.[9] However, beginning with the collapse of the Lisp Machine market in 1987, AI once again fell into disrepute, and a second, longer-lasting winter began.[11]

Up to this point, most of AI's funding had gone to projects that used high-level symbols to represent mental objects like plans, goals, beliefs, and known facts. In the 1980s, some researchers began to doubt that this approach would be able to imitate all the processes of human cognition, especially perception, robotics, learning and pattern recognition,[307] and began to look into "sub-symbolic" approaches.[308] Rodney Brooks rejected "representation" in general and focussed directly on engineering machines that move and survive.[w] Judea Pearl, Lofti Zadeh and others developed methods that handled incomplete and uncertain information by making reasonable guesses rather than precise logic.[90][313] But the most important development was the revival of "connectionism", including neural network research, by Geoffrey Hinton and others.[314] In 1990, Yann LeCun successfully showed that convolutional neural networks can recognize handwritten digits, the first of many successful applications of neural networks.[315]

AI gradually restored its reputation in the late 1990s and early 21st century by exploiting formal mathematical methods and by finding specific solutions to specific problems. This "narrow" and "formal" focus allowed researchers to produce verifiable results and collaborate with other fields (such as statistics, economics and mathematics).[316] By 2000, solutions developed by AI researchers were being widely used, although in the 1990s they were rarely described as "artificial intelligence".[317] However, several academic researchers became concerned that AI was no longer pursuing its original goal of creating versatile, fully intelligent machines. Beginning around 2002, they founded the subfield of artificial general intelligence (or "AGI"), which had several well-funded institutions by the 2010s.[15]

Deep learning began to dominate industry benchmarks in 2012 and was adopted throughout the field.[12] For many specific tasks, other methods were abandoned.[x] Deep learning's success was based on both hardware improvements (faster computers,[319] graphics processing units, cloud computing[320]) and access to large amounts of data[321] (including curated datasets,[320] such as ImageNet). Deep learning's success led to an enormous increase in interest and funding in AI.[y] The amount of machine learning research (measured by total publications) increased by 50% in the years 2015–2019.[280]

In 2016, issues of fairness and the misuse of technology were catapulted into center stage at machine learning conferences, publications vastly increased, funding became available, and many researchers re-focussed their careers on these issues. The alignment problem became a serious field of academic study.[257]

In the late teens and early 2020s, AGI companies began to deliver programs that created enormous interest. In 2015, AlphaGo, developed by DeepMind, beat the world champion Go player. The program was taught only the rules of the game and developed strategy by itself. GPT-3 is a large language model that was released in 2020 by OpenAI and is capable of generating high-quality human-like text.[322] These programs, and others, inspired an aggressive AI boom, where large companies began investing billions in AI research. According to AI Impacts, about $50 billion annually was invested in "AI" around 2022 in the U.S. alone and about 20% of the new U.S. Computer Science PhD graduates have specialized in "AI".[323] About 800,000 "AI"-related U.S. job openings existed in 2022.[324]

Philosophy

Defining artificial intelligence

Alan Turing wrote in 1950 "I propose to consider the question 'can machines think'?"[325] He advised changing the question from whether a machine "thinks", to "whether or not it is possible for machinery to show intelligent behaviour".[325] He devised the Turing test, which measures the ability of a machine to simulate human conversation.[295] Since we can only observe the behavior of the machine, it does not matter if it is "actually" thinking or literally has a "mind". Turing notes that we can not determine these things about other people but "it is usual to have a polite convention that everyone thinks."[326]

Russell and Norvig agree with Turing that intelligence must be defined in terms of external behavior, not internal structure.[1] However, they are critical that the test requires the machine to imitate humans. "Aeronautical engineering texts," they wrote, "do not define the goal of their field as making 'machines that fly so exactly like pigeons that they can fool other pigeons.'"[327] AI founder John McCarthy agreed, writing that "Artificial intelligence is not, by definition, simulation of human intelligence".[328]

McCarthy defines intelligence as "the computational part of the ability to achieve goals in the world".[329] Another AI founder, Marvin Minsky similarly describes it as "the ability to solve hard problems".[330] The leading AI textbook defines it as the study of agents that perceive their environment and take actions that maximize their chances of achieving defined goals.[1] These definitions view intelligence in terms of well-defined problems with well-defined solutions, where both the difficulty of the problem and the performance of the program are direct measures of the "intelligence" of the machine—and no other philosophical discussion is required, or may not even be possible.

Another definition has been adopted by Google,[331] a major practitioner in the field of AI. This definition stipulates the ability of systems to synthesize information as the manifestation of intelligence, similar to the way it is defined in biological intelligence.

Some authors have suggested in practice, that the definition of AI is vague and difficult to define, with contention as to whether classical algorithms should be categorised as AI,[332] with many companies during the early 2020s AI boom using the term as a marketing buzzword, often even if they did "not actually use AI in a material way".[333]

Evaluating approaches to AI

No established unifying theory or paradigm has guided AI research for most of its history.[z] The unprecedented success of statistical machine learning in the 2010s eclipsed all other approaches (so much so that some sources, especially in the business world, use the term "artificial intelligence" to mean "machine learning with neural networks"). This approach is mostly sub-symbolic, soft and narrow. Critics argue that these questions may have to be revisited by future generations of AI researchers.

Symbolic AI and its limits

Symbolic AI (or "GOFAI")[335] simulated the high-level conscious reasoning that people use when they solve puzzles, express legal reasoning and do mathematics. They were highly successful at "intelligent" tasks such as algebra or IQ tests. In the 1960s, Newell and Simon proposed the physical symbol systems hypothesis: "A physical symbol system has the necessary and sufficient means of general intelligent action."[336]

However, the symbolic approach failed on many tasks that humans solve easily, such as learning, recognizing an object or commonsense reasoning. Moravec's paradox is the discovery that high-level "intelligent" tasks were easy for AI, but low level "instinctive" tasks were extremely difficult.[337] Philosopher Hubert Dreyfus had argued since the 1960s that human expertise depends on unconscious instinct rather than conscious symbol manipulation, and on having a "feel" for the situation, rather than explicit symbolic knowledge.[338] Although his arguments had been ridiculed and ignored when they were first presented, eventually, AI research came to agree with him.[aa][20]

The issue is not resolved: sub-symbolic reasoning can make many of the same inscrutable mistakes that human intuition does, such as algorithmic bias. Critics such as Noam Chomsky argue continuing research into symbolic AI will still be necessary to attain general intelligence,[340][341] in part because sub-symbolic AI is a move away from explainable AI: it can be difficult or impossible to understand why a modern statistical AI program made a particular decision. The emerging field of neuro-symbolic artificial intelligence attempts to bridge the two approaches.

Neat vs. scruffy

"Neats" hope that intelligent behavior is described using simple, elegant principles (such as logic, optimization, or neural networks). "Scruffies" expect that it necessarily requires solving a large number of unrelated problems. Neats defend their programs with theoretical rigor, scruffies rely mainly on incremental testing to see if they work. This issue was actively discussed in the 1970s and 1980s,[342] but eventually was seen as irrelevant. Modern AI has elements of both.

Soft vs. hard computing

Finding a provably correct or optimal solution is intractable for many important problems.[19] Soft computing is a set of techniques, including genetic algorithms, fuzzy logic and neural networks, that are tolerant of imprecision, uncertainty, partial truth and approximation. Soft computing was introduced in the late 1980s and most successful AI programs in the 21st century are examples of soft computing with neural networks.

Narrow vs. general AI

AI researchers are divided as to whether to pursue the goals of artificial general intelligence and superintelligence directly or to solve as many specific problems as possible (narrow AI) in hopes these solutions will lead indirectly to the field's long-term goals.[343][344] General intelligence is difficult to define and difficult to measure, and modern AI has had more verifiable successes by focusing on specific problems with specific solutions. The experimental sub-field of artificial general intelligence studies this area exclusively.

Machine consciousness, sentience, and mind

The philosophy of mind does not know whether a machine can have a mind, consciousness and mental states, in the same sense that human beings do. This issue considers the internal experiences of the machine, rather than its external behavior. Mainstream AI research considers this issue irrelevant because it does not affect the goals of the field: to build machines that can solve problems using intelligence. Russell and Norvig add that "[t]he additional project of making a machine conscious in exactly the way humans are is not one that we are equipped to take on."[345] However, the question has become central to the philosophy of mind. It is also typically the central question at issue in artificial intelligence in fiction.

Consciousness

David Chalmers identified two problems in understanding the mind, which he named the "hard" and "easy" problems of consciousness.[346] The easy problem is understanding how the brain processes signals, makes plans and controls behavior. The hard problem is explaining how this feels or why it should feel like anything at all, assuming we are right in thinking that it truly does feel like something (Dennett's consciousness illusionism says this is an illusion). While human information processing is easy to explain, human subjective experience is difficult to explain. For example, it is easy to imagine a color-blind person who has learned to identify which objects in their field of view are red, but it is not clear what would be required for the person to know what red looks like.[347]

Computationalism and functionalism

Computationalism is the position in the philosophy of mind that the human mind is an information processing system and that thinking is a form of computing. Computationalism argues that the relationship between mind and body is similar or identical to the relationship between software and hardware and thus may be a solution to the mind–body problem. This philosophical position was inspired by the work of AI researchers and cognitive scientists in the 1960s and was originally proposed by philosophers Jerry Fodor and Hilary Putnam.[348]

Philosopher John Searle characterized this position as "strong AI": "The appropriately programmed computer with the right inputs and outputs would thereby have a mind in exactly the same sense human beings have minds."[ab] Searle counters this assertion with his Chinese room argument, which attempts to show that, even if a machine perfectly simulates human behavior, there is still no reason to suppose it also has a mind.[352]

AI welfare and rights

It is difficult or impossible to reliably evaluate whether an advanced AI is sentient (has the ability to feel), and if so, to what degree.[353] But if there is a significant chance that a given machine can feel and suffer, then it may be entitled to certain rights or welfare protection measures, similarly to animals.[354][355] Sapience (a set of capacities related to high intelligence, such as discernment or self-awareness) may provide another moral basis for AI rights.[354] Robot rights are also sometimes proposed as a practical way to integrate autonomous agents into society.[356]

In 2017, the European Union considered granting "electronic personhood" to some of the most capable AI systems. Similarly to the legal status of companies, it would have conferred rights but also responsibilities.[357] Critics argued in 2018 that granting rights to AI systems would downplay the importance of human rights, and that legislation should focus on user needs rather than speculative futuristic scenarios. They also noted that robots lacked the autonomy to take part to society on their own.[358][359]

Progress in AI increased interest in the topic. Proponents of AI welfare and rights often argue that AI sentience, if it emerges, would be particularly easy to deny. They warn that this may be a moral blind spot analogous to slavery or factory farming, which could lead to large-scale suffering if sentient AI is created and carelessly exploited.[355][354]

Future

Superintelligence and the singularity

A superintelligence is a hypothetical agent that would possess intelligence far surpassing that of the brightest and most gifted human mind.[344]

If research into artificial general intelligence produced sufficiently intelligent software, it might be able to reprogram and improve itself. The improved software would be even better at improving itself, leading to what I. J. Good called an "intelligence explosion" and Vernor Vinge called a "singularity".[360]

However, technologies cannot improve exponentially indefinitely, and typically follow an S-shaped curve, slowing when they reach the physical limits of what the technology can do.[361]

Transhumanism

Robot designer Hans Moravec, cyberneticist Kevin Warwick, and inventor Ray Kurzweil have predicted that humans and machines will merge in the future into cyborgs that are more capable and powerful than either. This idea, called transhumanism, has roots in Aldous Huxley and Robert Ettinger.[362]

Edward Fredkin argues that "artificial intelligence is the next stage in evolution", an idea first proposed by Samuel Butler's "Darwin among the Machines" as far back as 1863, and expanded upon by George Dyson in his 1998 book Darwin Among the Machines: The Evolution of Global Intelligence.[363]

In fiction

The word "robot" itself was coined by Karel Čapek in his 1921 play R.U.R., the title standing for "Rossum's Universal Robots"

Thought-capable artificial beings have appeared as storytelling devices since antiquity,[364] and have been a persistent theme in science fiction.[365]

A common trope in these works began with Mary Shelley's Frankenstein, where a human creation becomes a threat to its masters. This includes such works as Arthur C. Clarke's and Stanley Kubrick's 2001: A Space Odyssey (both 1968), with HAL 9000, the murderous computer in charge of the Discovery One spaceship, as well as The Terminator (1984) and The Matrix (1999). In contrast, the rare loyal robots such as Gort from The Day the Earth Stood Still (1951) and Bishop from Aliens (1986) are less prominent in popular culture.[366]

Isaac Asimov introduced the Three Laws of Robotics in many stories, most notably with the "Multivac" super-intelligent computer. Asimov's laws are often brought up during lay discussions of machine ethics;[367] while almost all artificial intelligence researchers are familiar with Asimov's laws through popular culture, they generally consider the laws useless for many reasons, one of which is their ambiguity.[368]

Several works use AI to force us to confront the fundamental question of what makes us human, showing us artificial beings that have the ability to feel, and thus to suffer. This appears in Karel Čapek's R.U.R., the films A.I. Artificial Intelligence and Ex Machina, as well as the novel Do Androids Dream of Electric Sheep?, by Philip K. Dick. Dick considers the idea that our understanding of human subjectivity is altered by technology created with artificial intelligence.[369]

See also

Explanatory notes

  1. ^ Jump up to: a b This list of intelligent traits is based on the topics covered by the major AI textbooks, including: Russell & Norvig (2021), Luger & Stubblefield (2004), Poole, Mackworth & Goebel (1998) and Nilsson (1998)
  2. ^ Jump up to: a b This list of tools is based on the topics covered by the major AI textbooks, including: Russell & Norvig (2021), Luger & Stubblefield (2004), Poole, Mackworth & Goebel (1998) and Nilsson (1998)
  3. ^ It is among the reasons that expert systems proved to be inefficient for capturing knowledge.[34][35]
  4. ^ "Rational agent" is general term used in economics, philosophy and theoretical artificial intelligence. It can refer to anything that directs its behavior to accomplish goals, such as a person, an animal, a corporation, a nation, or in the case of AI, a computer program.
  5. ^ Alan Turing discussed the centrality of learning as early as 1950, in his classic paper "Computing Machinery and Intelligence".[46] In 1956, at the original Dartmouth AI summer conference, Ray Solomonoff wrote a report on unsupervised probabilistic machine learning: "An Inductive Inference Machine".[47]
  6. ^ See AI winter § Machine translation and the ALPAC report of 1966
  7. ^ Compared with symbolic logic, formal Bayesian inference is computationally expensive. For inference to be tractable, most observations must be conditionally independent of one another. AdSense uses a Bayesian network with over 300 million edges to learn which ads to serve.[97]
  8. ^ Expectation–maximization, one of the most popular algorithms in machine learning, allows clustering in the presence of unknown latent variables.[99]
  9. ^ Some form of deep neural networks (without a specific learning algorithm) were described by: Warren S. McCulloch and Walter Pitts (1943)[118] Alan Turing (1948);[119] Karl Steinbuch and Roger David Joseph (1961).[120] Deep or recurrent networks that learned (or used gradient descent) were developed by: Frank Rosenblatt(1957);[119] Oliver Selfridge (1959);[120] Alexey Ivakhnenko and Valentin Lapa (1965);[121] Kaoru Nakano (1971);[122] Shun-Ichi Amari (1972);[122] John Joseph Hopfield (1982).[122] Precursors to backpropagation were developed by: Henry J. Kelley (1960);[119] Arthur E. Bryson (1962);[119] Stuart Dreyfus (1962);[119] Arthur E. Bryson and Yu-Chi Ho (1969);[119] Backpropagation was independently developed by: Seppo Linnainmaa (1970);[123] Paul Werbos (1974).[119]
  10. ^ Geoffrey Hinton said, of his work on neural networks in the 1990s, "our labeled datasets were thousands of times too small. [And] our computers were millions of times too slow"[124]
  11. ^ Including Jon Kleinberg (Cornell University), Sendhil Mullainathan (University of Chicago), Cynthia Chouldechova (Carnegie Mellon) and Sam Corbett-Davis (Stanford)[202]
  12. ^ Moritz Hardt (a director at the Max Planck Institute for Intelligent Systems) argues that machine learning "is fundamentally the wrong tool for a lot of domains, where you're trying to design interventions and mechanisms that change the world."[207]
  13. ^ When the law was passed in 2018, it still contained a form of this provision.
  14. ^ This is the United Nations' definition, and includes things like land mines as well.[221]
  15. ^ See table 4; 9% is both the OECD average and the U.S. average.[232]
  16. ^ Sometimes called a "robopocalypse"[240]
  17. ^ "Electronic brain" was the term used by the press around this time.[292][293]
  18. ^ Daniel Crevier wrote, "the conference is generally recognized as the official birthdate of the new science."[296] Russell and Norvig called the conference "the inception of artificial intelligence."[118]
  19. ^ Russell and Norvig wrote "for the next 20 years the field would be dominated by these people and their students."[297]
  20. ^ Russell and Norvig wrote "it was astonishing whenever a computer did anything kind of smartish".[298]
  21. ^ The programs described are Arthur Samuel's checkers program for the IBM 701, Daniel Bobrow's STUDENT, Newell and Simon's Logic Theorist and Terry Winograd's SHRDLU.
  22. ^ Russell and Norvig write: "in almost all cases, these early systems failed on more difficult problems"[302]
  23. ^ Embodied approaches to AI[309] were championed by Hans Moravec[310] and Rodney Brooks[311] and went by many names: Nouvelle AI.[311] Developmental robotics.[312]
  24. ^ Matteo Wong wrote in The Atlantic: "Whereas for decades, computer-science fields such as natural-language processing, computer vision, and robotics used extremely different methods, now they all use a programming method called "deep learning." As a result, their code and approaches have become more similar, and their models are easier to integrate into one another."[318]
  25. ^ Jack Clark wrote in Bloomberg: "After a half-decade of quiet breakthroughs in artificial intelligence, 2015 has been a landmark year. Computers are smarter and learning faster than ever", and noted that the number of software projects that use machine learning at Google increased from a "sporadic usage" in 2012 to more than 2,700 projects in 2015.[320]
  26. ^ Nils Nilsson wrote in 1983: "Simply put, there is wide disagreement in the field about what AI is all about."[334]
  27. ^ Daniel Crevier wrote that "time has proven the accuracy and perceptiveness of some of Dreyfus's comments. Had he formulated them less aggressively, constructive actions they suggested might have been taken much earlier."[339]
  28. ^ Searle presented this definition of "Strong AI" in 1999.[349] Searle's original formulation was "The appropriately programmed computer really is a mind, in the sense that computers given the right programs can be literally said to understand and have other cognitive states."[350] Strong AI is defined similarly by Russell and Norvig: "Stong AI – the assertion that machines that do so are actually thinking (as opposed to simulating thinking)."[351]

References

  1. ^ Jump up to: a b c Russell & Norvig (2021), pp. 1–4.
  2. ^ AI set to exceed human brain power Archived 2008-02-19 at the Wayback Machine CNN.com (July 26, 2006)
  3. ^ Kaplan, Andreas; Haenlein, Michael (2019). "Siri, Siri, in my hand: Who's the fairest in the land? On the interpretations, illustrations, and implications of artificial intelligence". Business Horizons. 62: 15–25. doi:10.1016/j.bushor.2018.08.004. ISSN 0007-6813. S2CID 158433736.
  4. ^ Jump up to: a b c d Copeland, J., ed. (2004). The Essential Turing: the ideas that gave birth to the computer age. Oxford, England: Clarendon Press. ISBN 0-19-825079-7.
  5. ^ Jump up to: a b Dartmouth workshop: The proposal:
  6. ^ Kaplan, Andreas (2022). Artificial Intelligence, Business and Civilization: Our Fate Made in Machines. Routledge focus on business and management. New York, NY: Routledge. ISBN 978-1-000-56333-7.
  7. ^ Marquis, Pierre; Papini, Odile; Prade, Henri, eds. (2020). A Guided Tour of Artificial Intelligence Research: Volume III: Interfaces and Applications of Artificial Intelligence. Cham: Springer International Publishing. pp. xiii. doi:10.1007/978-3-030-06170-8. ISBN 978-3-030-06169-2.
  8. ^ Jump up to: a b Successful programs the 1960s:
  9. ^ Jump up to: a b Funding initiatives in the early 1980s: Fifth Generation Project (Japan), Alvey (UK), Microelectronics and Computer Technology Corporation (US), Strategic Computing Initiative (US):
  10. ^ Jump up to: a b First AI Winter, Lighthill report, Mansfield Amendment
  11. ^ Jump up to: a b Second AI Winter:
  12. ^ Jump up to: a b Deep learning revolution, AlexNet:
  13. ^ Toews (2023).
  14. ^ Frank (2023).
  15. ^ Jump up to: a b c Artificial general intelligence: Proposal for the modern version: Warnings of overspecialization in AI from leading researchers:
  16. ^ Russell & Norvig (2021, §1.2).
  17. ^ Problem-solving, puzzle solving, game playing, and deduction:
  18. ^ Uncertain reasoning:
  19. ^ Jump up to: a b c Intractability and efficiency and the combinatorial explosion:
  20. ^ Jump up to: a b c Psychological evidence of the prevalence of sub-symbolic reasoning and knowledge:
  21. ^ Knowledge representation and knowledge engineering:
  22. ^ Smoliar & Zhang (1994).
  23. ^ Neumann & Möller (2008).
  24. ^ Kuperman, Reichley & Bailey (2006).
  25. ^ McGarry (2005).
  26. ^ Bertini, Del Bimbo & Torniai (2006).
  27. ^ Russell & Norvig (2021), pp. 272.
  28. ^ Representing categories and relations: Semantic networks, description logics, inheritance (including frames, and scripts):
  29. ^ Representing events and time:Situation calculus, event calculus, fluent calculus (including solving the frame problem):
  30. ^ Causal calculus:
  31. ^ Representing knowledge about knowledge: Belief calculus, modal logics:
  32. ^ Jump up to: a b Default reasoning, Frame problem, default logic, non-monotonic logics, circumscription, closed world assumption, abduction: (Poole et al. places abduction under "default reasoning". Luger et al. places this under "uncertain reasoning").
  33. ^ Jump up to: a b Breadth of commonsense knowledge:
  34. ^ Newquist (1994), p. 296.
  35. ^ Crevier (1993), pp. 204–208.
  36. ^ Russell & Norvig (2021), p. 528.
  37. ^ Automated planning:
  38. ^ Automated decision making, Decision theory:
  39. ^ Classical planning:
  40. ^ Sensorless or "conformant" planning, contingent planning, replanning (a.k.a online planning):
  41. ^ Uncertain preferences: Inverse reinforcement learning:
  42. ^ Information value theory:
  43. ^ Markov decision process:
  44. ^ Game theory and multi-agent decision theory:
  45. ^ Learning:
  46. ^ Turing (1950).
  47. ^ Solomonoff (1956).
  48. ^ Unsupervised learning:
  49. ^ Jump up to: a b Supervised learning:
  50. ^ Reinforcement learning:
  51. ^ Transfer learning:
  52. ^ "Artificial Intelligence (AI): What Is AI and How Does It Work? | Built In". builtin.com. Retrieved 30 October 2023.
  53. ^ Computational learning theory:
  54. ^ Natural language processing (NLP):
  55. ^ Subproblems of NLP:
  56. ^ Russell & Norvig (2021), pp. 856–858.
  57. ^ Dickson (2022).
  58. ^ Modern statistical and deep learning approaches to NLP:
  59. ^ Vincent (2019).
  60. ^ Russell & Norvig (2021), pp. 875–878.
  61. ^ Bushwick (2023).
  62. ^ Computer vision:
  63. ^ Russell & Norvig (2021), pp. 849–850.
  64. ^ Russell & Norvig (2021), pp. 895–899.
  65. ^ Russell & Norvig (2021), pp. 899–901.
  66. ^ Challa et al. (2011).
  67. ^ Russell & Norvig (2021), pp. 931–938.
  68. ^ MIT AIL (2014).
  69. ^ Affective computing:
  70. ^ Waddell (2018).
  71. ^ Poria et al. (2017).
  72. ^ Search algorithms:
  73. ^ State space search:
  74. ^ Russell & Norvig (2021), §11.2.
  75. ^ Uninformed searches (breadth first search, depth-first search and general state space search):
  76. ^ Heuristic or informed searches (e.g., greedy best first and A*):
  77. ^ Adversarial search:
  78. ^ Local or "optimization" search:
  79. ^ Singh Chauhan, Nagesh (18 December 2020). "Optimization Algorithms in Neural Networks". KDnuggets. Retrieved 13 January 2024.
  80. ^ Evolutionary computation:
  81. ^ Merkle & Middendorf (2013).
  82. ^ Logic:
  83. ^ Propositional logic:
  84. ^ First-order logic and features such as equality:
  85. ^ Logical inference:
  86. ^ logical deduction as search:
  87. ^ Resolution and unification:
  88. ^ Warren, D.H.; Pereira, L.M.; Pereira, F. (1977). "Prolog-the language and its implementation compared with Lisp". ACM SIGPLAN Notices. 12 (8): 109–115. doi:10.1145/872734.806939.
  89. ^ Fuzzy logic:
  90. ^ Jump up to: a b Stochastic methods for uncertain reasoning:
  91. ^ decision theory and decision analysis:
  92. ^ Information value theory:
  93. ^ Markov decision processes and dynamic decision networks:
  94. ^ Jump up to: a b c Стохастические временные модели: Hidden Markov model: Kalman filters: Динамические байесовские сети :
  95. ^ Теория игр и проектирование механизмов :
  96. ^ Байесовские сети :
  97. ^ Домингос (2015) , глава 6.
  98. ^ байесовского вывода Алгоритм :
  99. ^ Домингос (2015) , с. 210.
  100. ^ Байесовское обучение и алгоритм максимизации ожидания :
  101. ^ Байесовская теория принятия решений и байесовские сети принятия решений :
  102. ^ Статистические методы обучения и классификаторы :
  103. ^ Деревья решений :
  104. ^ Непараметрические модели обучения, такие как K-ближайший сосед и машины опорных векторов :
  105. ^ Домингос (2015) , с. 152.
  106. ^ Наивный байесовский классификатор :
  107. ^ Jump up to: а б Нейронные сети:
  108. ^ Вычисление градиента в вычислительных графах, обратное распространение ошибки , автоматическое дифференцирование :
  109. ^ Теорема об универсальной аппроксимации : Теорема:
  110. ^ Нейронные сети прямого распространения :
  111. ^ Рекуррентные нейронные сети :
  112. ^ Перцептроны :
  113. ^ Jump up to: а б Глубокое обучение :
  114. ^ Сверточные нейронные сети :
  115. ^ Дэн и Ю (2014) , стр. 199–200.
  116. ^ Чиресан, Мейер и Шмидхубер (2012) .
  117. ^ Рассел и Норвиг (2021) , с. 751.
  118. ^ Jump up to: а б с Рассел и Норвиг (2021) , с. 17.
  119. ^ Jump up to: а б с д и ж г Рассел и Норвиг (2021) , с. 785.
  120. ^ Jump up to: а б Шмидхубер (2022) , §5.
  121. ^ Шмидхубер (2022) , §6.
  122. ^ Jump up to: а б с Шмидхубер (2022) , §7.
  123. ^ Шмидхубер (2022) , §8.
  124. ^ Цитируется по Кристиану (2020 , стр. 22).
  125. ^ Смит (2023) .
  126. ^ «Объяснение: Генеративный ИИ» . 9 ноября 2023 г.
  127. ^ «Инструменты для написания искусственного интеллекта и создания контента» . Технологии преподавания и обучения Массачусетского технологического института Слоана . Проверено 25 декабря 2023 г.
  128. ^ Мармуйе (2023) .
  129. ^ Кобелус (2019) .
  130. ^ Томасон, Джеймс (21 мая 2024 г.). «Восстание Mojo: возрождение языков программирования, ориентированных на искусственный интеллект» . ВенчурБит . Проверено 26 мая 2024 г.
  131. ^ Водецки, Бен (5 мая 2023 г.). «7 языков программирования искусственного интеллекта, которые вам нужно знать» . ИИ-бизнес .
  132. ^ Давенпорт, Т; Калакота, Р. (июнь 2019 г.). «Потенциал искусственного интеллекта в здравоохранении» . Будущее Здоровьеc Дж . 6 (2): 94–98. doi : 10.7861/futurehosp.6-2-94 . ПМК   6616181 . ПМИД   31363513 .
  133. ^ Jump up to: а б Бакс, Моник; Торп, Джордан; Романов, Валентин (декабрь 2023 г.). «Будущее персонализированной сердечно-сосудистой медицины требует 3D- и 4D-печати, стволовых клеток и искусственного интеллекта» . Границы в сенсорах . 4 . дои : 10.3389/fsens.2023.1294721 . ISSN   2673-5067 .
  134. ^ Джампер, Дж; Эванс, Р; Притцель, А (2021). «Высокоточное предсказание структуры белка с помощью AlphaFold» . Природа . 596 (7873): 583–589. Бибкод : 2021Natur.596..583J . дои : 10.1038/s41586-021-03819-2 . ПМЦ   8371605 . PMID   34265844 .
  135. ^ «ИИ открывает новый класс антибиотиков для уничтожения устойчивых к лекарствам бактерий» . 20 декабря 2023 г.
  136. ^ «ИИ ускоряет разработку лекарств от болезни Паркинсона в десять раз» . Кембриджский университет. 17 апреля 2024 г.
  137. ^ Хорн, Роберт И.; Анджеевска, Ева А.; Алам, Парвез; Бротзакис, З. Фейдон; Шривастава, Анкит; Обер, Алиса; Новинска, Магдалена; Грегори, Ребекка С.; Стаатс, Роксин; Поссенти, Андреа; Чиа, Шон; Сорманни, Пьетро; Гетти, Бернардино; Коги, Байрон; Ноулз, Туомас П.Дж.; Вендруколо, Микеле (17 апреля 2024 г.). «Открытие мощных ингибиторов агрегации α-синуклеина с использованием структурного итеративного обучения» . Химическая биология природы . 20 (5). Природа: 634–645. дои : 10.1038/s41589-024-01580-x . ПМЦ   11062903 . ПМИД   38632492 .
  138. ^ Грант, Юджин Ф.; Ларднер, Рекс (25 июля 1952 г.). «Городской разговор - Это» . Житель Нью-Йорка . ISSN   0028-792X . Проверено 28 января 2024 г.
  139. ^ Андерсон, Марк Роберт (11 мая 2017 г.). «Двадцать лет спустя после матча Deep Blue против Каспарова: как шахматный матч положил начало революции больших данных» . Разговор . Проверено 28 января 2024 г.
  140. ^ Маркофф, Джон (16 февраля 2011 г.). «Компьютер побеждает в игре «Jeopardy!»: это не тривиально» . Нью-Йорк Таймс . ISSN   0362-4331 . Проверено 28 января 2024 г.
  141. ^ Байфорд, Сэм (27 мая 2017 г.). «AlphaGo уходит из соревновательного го после победы над номером один в мире со счетом 3–0» . Грань . Проверено 28 января 2024 г.
  142. ^ Браун, Ноам; Сандхольм, Туомас (30 августа 2019 г.). «Сверхчеловеческий ИИ для многопользовательского покера» . Наука . 365 (6456): 885–890. Бибкод : 2019Sci...365..885B . дои : 10.1126/science.aay2400 . ISSN   0036-8075 . ПМИД   31296650 .
  143. ^ «MuZero: Освоение го, шахмат, сёги и Atari без правил» . Гугл ДипМайнд . 23 декабря 2020 г. Проверено 28 января 2024 г.
  144. ^ Сэмпл, Ян (30 октября 2019 г.). «ИИ становится гроссмейстером в «чертовски сложном» StarCraft II» . Хранитель . ISSN   0261-3077 . Проверено 28 января 2024 г.
  145. ^ Вурман, PR; Барретт, С.; Кавамото, К. (2022). «Обгоняйте водителей-чемпионов Gran Turismo с помощью глубокого обучения с подкреплением». Природа . 602 (7896): 223–228. Бибкод : 2022Natur.602..223W . дои : 10.1038/s41586-021-04357-7 . ПМИД   35140384 .
  146. ^ Уилкинс, Алекс (13 марта 2024 г.). «ИИ Google учится играть в видеоигры с открытым миром, наблюдая за ними» . Новый учёный . Проверено 21 июля 2024 г.
  147. ^ Уэсато, Дж. и др.: Улучшение математических рассуждений с помощью контроля процесса. openai.com, 31 мая 2023 г. Проверено 7 августа 2024 г.
  148. ^ Робертс, Шивон (25 июля 2024 г.). «ИИ получил серебряную медаль за решение задач Международной математической олимпиады» . nytimes.com . Проверено 7 августа 2024 г.
  149. ^ ЛЛЕММА . eleuther.ai. Проверено 7 августа 2024 г.
  150. ^ AI Математика. Caesars Labs, 2024 г. Проверено 7 августа 2024 г.
  151. ^ Алекс МакФарланд: 7 лучших ИИ для математических инструментов. unite.ai. Проверено 7 августа 2024 г.
  152. ^ Мэтью Финио и Аманда Дауни: Учебник IBM Think 2024, «Что такое искусственный интеллект (ИИ) в финансах?» 8 декабря 2023 г.
  153. ^ М. Николас Дж. Фирзли: Pensions Age/журнал European Pensions, «Искусственный интеллект: спросите отрасль», май, июнь 2024 г., https://videovoice.org/ai-in-finance-innovation-entrepreneurship-vs-over-regulation- с-искусственным-интеллектуальным-искусственным-исследованием-не будет-работать-как-задумано/
  154. ^ Jump up to: a b c Congressional Research Service (2019). Artificial Intelligence and National Security (PDF). Washington, DC: Congressional Research Service.PD-notice
  155. ^ Jump up to: a b Slyusar, Vadym (2019). "Artificial intelligence as the basis of future control networks". ResearchGate. doi:10.13140/RG.2.2.30247.50087.
  156. ^ Knight, Will. "The US and 30 Other Nations Agree to Set Guardrails for Military AI". Wired. ISSN 1059-1028. Retrieved 24 January 2024.
  157. ^ Marcelline, Marco (27 May 2023). "ChatGPT: Most Americans Know About It, But Few Actually Use the AI Chatbot". PCMag. Retrieved 28 January 2024.
  158. ^ Lu, Donna (31 March 2023). "Misinformation, mistakes and the Pope in a puffer: what rapidly evolving AI can – and can't – do". The Guardian. ISSN 0261-3077. Retrieved 28 January 2024.
  159. ^ Hurst, Luke (23 May 2023). "How a fake image of a Pentagon explosion shared on Twitter caused a real dip on Wall Street". euronews. Retrieved 28 January 2024.
  160. ^ Ransbotham, Sam; Kiron, David; Gerbert, Philipp; Reeves, Martin (6 September 2017). "Reshaping Business With Artificial Intelligence". MIT Sloan Management Review. Archived from the original on 13 February 2024.
  161. ^ Sun, Yuran; Zhao, Xilei; Lovreglio, Ruggiero; Kuligowski, Erica (1 January 2024), Naser, M. Z. (ed.), "8 - AI for large-scale evacuation modeling: promises and challenges", Interpretable Machine Learning for the Analysis, Design, Assessment, and Informed Decision Making for Civil Infrastructure, Woodhead Publishing Series in Civil and Structural Engineering, Woodhead Publishing, pp. 185–204, ISBN 978-0-12-824073-1, retrieved 28 June 2024
  162. ^ Gomaa, Islam; Adelzadeh, Masoud; Gwynne, Steven; Spencer, Bruce; Ko, Yoon; Bénichou, Noureddine; Ma, Chunyun; Elsagan, Nour; Duong, Dana; Zalok, Ehab; Kinateder, Max (1 November 2021). "A Framework for Intelligent Fire Detection and Evacuation System". Fire Technology. 57 (6): 3179–3185. doi:10.1007/s10694-021-01157-3. ISSN 1572-8099.
  163. ^ Zhao, Xilei; Lovreglio, Ruggiero; Nilsson, Daniel (1 May 2020). "Modelling and interpreting pre-evacuation decision-making using machine learning". Automation in Construction. 113: 103140. doi:10.1016/j.autcon.2020.103140. ISSN 0926-5805.
  164. ^ Simonite (2016).
  165. ^ Russell & Norvig (2021), p. 987.
  166. ^ Laskowski (2023).
  167. ^ GAO (2022).
  168. ^ Valinsky (2019).
  169. ^ Russell & Norvig (2021), p. 991.
  170. ^ Russell & Norvig (2021), pp. 991–992.
  171. ^ Christian (2020), p. 63.
  172. ^ Vincent (2022).
  173. ^ Kopel, Matthew. "Copyright Services: Fair Use". Cornell University Library. Retrieved 26 April 2024.
  174. ^ Burgess, Matt. "How to Stop Your Data From Being Used to Train AI". Wired. ISSN 1059-1028. Retrieved 26 April 2024.
  175. ^ Reisner (2023).
  176. ^ Alter & Harris (2023).
  177. ^ "Getting the Innovation Ecosystem Ready for AI. An IP policy toolkit" (PDF). WIPO.
  178. ^ Hammond, George (27 December 2023). "Big Tech is spending more than VC firms on AI startups". Ars Technica. Archived from the original on 10 January 2024.
  179. ^ Wong, Matteo (24 October 2023). "The Future of AI Is GOMA". The Atlantic. Archived from the original on 5 January 2024.
  180. ^ "Big tech and the pursuit of AI dominance". The Economist. 26 March 2023. Archived from the original on 29 December 2023.
  181. ^ Fung, Brian (19 December 2023). "Where the battle to dominate AI may be won". CNN Business. Archived from the original on 13 January 2024.
  182. ^ Metz, Cade (5 July 2023). "In the Age of A.I., Tech's Little Guys Need Big Friends". The New York Times.
  183. ^ "Electricity 2024 – Analysis". IEA. 24 January 2024. Retrieved 13 July 2024.
  184. ^ Calvert, Brian (28 March 2024). "AI already uses as much energy as a small country. It's only the beginning". Vox. New York, NY.
  185. ^ Halper, Evan; O'Donovan, Caroline (21 June 2024). "AI is exhausting the power grid. Tech firms are seeking a miracle solution". Washington Post.
  186. ^ Davenport, Carly. "AI Data Centers and the Coming YS Power Demand Surge" (PDF). Goldman Sachs.
  187. ^ Ryan, Carol (12 April 2024). "Energy-Guzzling AI Is Also the Future of Energy Savings". Wall Street Journal. Dow Jones.
  188. ^ Hiller, Jennifer (1 July 2024). "Tech Industry Wants to Lock Up Nuclear Power for AI". Wall Street Journal. Dow Jones.
  189. ^ Nicas (2018).
  190. ^ Rainie, Lee; Keeter, Scott; Perrin, Andrew (22 July 2019). "Trust and Distrust in America". Pew Research Center. Archived from the original on 22 February 2024.
  191. ^ Williams (2023).
  192. ^ Taylor & Hern (2023).
  193. ^ Jump up to: a b Samuel, Sigal (19 April 2022). "Why it's so damn hard to make AI fair and unbiased". Vox. Retrieved 24 July 2024.
  194. ^ Jump up to: a b Rose (2023).
  195. ^ CNA (2019).
  196. ^ Goffrey (2008), p. 17.
  197. ^ Berdahl et al. (2023); Goffrey (2008, p. 17); Rose (2023); Russell & Norvig (2021, p. 995)
  198. ^ Christian (2020), p. 25.
  199. ^ Jump up to: a b Russell & Norvig (2021), p. 995.
  200. ^ Grant & Hill (2023).
  201. ^ Larson & Angwin (2016).
  202. ^ Christian (2020), p. 67–70.
  203. ^ Christian (2020, pp. 67–70); Russell & Norvig (2021, pp. 993–994)
  204. ^ Russell & Norvig (2021, p. 995); Lipartito (2011, p. 36); Goodman & Flaxman (2017, p. 6); Christian (2020, pp. 39–40, 65)
  205. ^ Quoted in Christian (2020, p. 65).
  206. ^ Russell & Norvig (2021, p. 994); Christian (2020, pp. 40, 80–81)
  207. ^ Quoted in Christian (2020, p. 80)
  208. ^ Dockrill (2022).
  209. ^ Sample (2017).
  210. ^ "Black Box AI". 16 June 2023.
  211. ^ Christian (2020), p. 110.
  212. ^ Christian (2020), pp. 88–91.
  213. ^ Christian (2020, p. 83); Russell & Norvig (2021, p. 997)
  214. ^ Christian (2020), p. 91.
  215. ^ Christian (2020), p. 83.
  216. ^ Verma (2021).
  217. ^ Rothman (2020).
  218. ^ Christian (2020), pp. 105–108.
  219. ^ Christian (2020), pp. 108–112.
  220. ^ Ropek, Lucas (21 May 2024). "New Anthropic Research Sheds Light on AI's 'Black Box'". Gizmodo. Retrieved 23 May 2024.
  221. ^ Russell & Norvig (2021), p. 989.
  222. ^ Jump up to: a b Russell & Norvig (2021), pp. 987–990.
  223. ^ Russell & Norvig (2021), p. 988.
  224. ^ Robitzski (2018); Sainato (2015)
  225. ^ Harari (2018).
  226. ^ Buckley, Chris; Mozur, Paul (22 May 2019). "How China Uses High-Tech Surveillance to Subdue Minorities". The New York Times.
  227. ^ "Security lapse exposed a Chinese smart city surveillance system". 3 May 2019. Archived from the original on 7 March 2021. Retrieved 14 September 2020.
  228. ^ Urbina et al. (2022).
  229. ^ Jump up to: a b E McGaughey, 'Will Robots Automate Your Job Away? Full Employment, Basic Income, and Economic Democracy' (2022) 51(3) Industrial Law Journal 511–559 Archived 27 May 2023 at the Wayback Machine
  230. ^ Ford & Colvin (2015);McGaughey (2022)
  231. ^ IGM Chicago (2017).
  232. ^ Arntz, Gregory & Zierahn (2016), p. 33.
  233. ^ Lohr (2017); Frey & Osborne (2017); Arntz, Gregory & Zierahn (2016, p. 33)
  234. ^ Zhou, Viola (11 April 2023). "AI is already taking video game illustrators' jobs in China". Rest of World. Retrieved 17 August 2023.
  235. ^ Carter, Justin (11 April 2023). "China's game art industry reportedly decimated by growing AI use". Game Developer. Retrieved 17 August 2023.
  236. ^ Morgenstern (2015).
  237. ^ Mahdawi (2017); Thompson (2014)
  238. ^ Tarnoff, Ben (4 August 2023). "Lessons from Eliza". The Guardian Weekly. pp. 34–39.
  239. ^ Cellan-Jones (2014).
  240. ^ Russell & Norvig 2021, p. 1001.
  241. ^ Bostrom (2014).
  242. ^ Russell (2019).
  243. ^ Bostrom (2014); Müller & Bostrom (2014); Bostrom (2015).
  244. ^ Harari (2023).
  245. ^ Müller & Bostrom (2014).
  246. ^ Leaders' concerns about the existential risks of AI around 2015:
  247. ^ ""Godfather of artificial intelligence" talks impact and potential of new AI". CBS News. 25 March 2023. Archived from the original on 28 March 2023. Retrieved 28 March 2023.
  248. ^ Pittis, Don (4 May 2023). "Canadian artificial intelligence leader Geoffrey Hinton piles on fears of computer takeover". CBC.
  249. ^ "'50-50 chance' that AI outsmarts humanity, Geoffrey Hinton says - BNN Bloomberg". Bloomberg BNN. 14 June 2024. Retrieved 6 July 2024.
  250. ^ Valance (2023).
  251. ^ Taylor, Josh (7 May 2023). "Rise of artificial intelligence is inevitable but should not be feared, 'father of AI' says". The Guardian. Retrieved 26 May 2023.
  252. ^ Colton, Emma (7 May 2023). "'Father of AI' says tech fears misplaced: 'You cannot stop it'". Fox News. Retrieved 26 May 2023.
  253. ^ Jones, Hessie (23 May 2023). "Juergen Schmidhuber, Renowned 'Father Of Modern AI,' Says His Life's Work Won't Lead To Dystopia". Forbes. Retrieved 26 May 2023.
  254. ^ McMorrow, Ryan (19 December 2023). "Andrew Ng: 'Do we think the world is better off with more or less intelligence?'". Financial Times. Retrieved 30 December 2023.
  255. ^ Levy, Steven (22 December 2023). "How Not to Be Stupid About AI, With Yann LeCun". Wired. Retrieved 30 December 2023.
  256. ^ Arguments that AI is not an imminent risk:
  257. ^ Jump up to: a b Christian (2020), pp. 67, 73.
  258. ^ Yudkowsky (2008).
  259. ^ Jump up to: a b Anderson & Anderson (2011).
  260. ^ AAAI (2014).
  261. ^ Wallach (2010).
  262. ^ Russell (2019), p. 173.
  263. ^ Stewart, Ashley; Melton, Monica. "Hugging Face CEO says he's focused on building a 'sustainable model' for the $4.5 billion open-source-AI startup". Business Insider. Retrieved 14 April 2024.
  264. ^ Wiggers, Kyle (9 April 2024). "Google open sources tools to support AI model development". TechCrunch. Retrieved 14 April 2024.
  265. ^ Heaven, Will Douglas (12 May 2023). "The open-source AI boom is built on Big Tech's handouts. How long will it last?". MIT Technology Review. Retrieved 14 April 2024.
  266. ^ Brodsky, Sascha (19 December 2023). "Mistral AI's New Language Model Aims for Open Source Supremacy". AI Business.
  267. ^ Edwards, Benj (22 February 2024). "Stability announces Stable Diffusion 3, a next-gen AI image generator". Ars Technica. Retrieved 14 April 2024.
  268. ^ Marshall, Matt (29 January 2024). "How enterprises are using open source LLMs: 16 examples". VentureBeat.
  269. ^ Piper, Kelsey (2 February 2024). "Should we make our most powerful AI models open source to all?". Vox. Retrieved 14 April 2024.
  270. ^ Alan Turing Institute (2019). "Understanding artificial intelligence ethics and safety" (PDF).
  271. ^ Alan Turing Institute (2023). "AI Ethics and Governance in Practice" (PDF).
  272. ^ Floridi, Luciano; Cowls, Josh (23 June 2019). "A Unified Framework of Five Principles for AI in Society". Harvard Data Science Review. 1 (1). doi:10.1162/99608f92.8cd550d1. S2CID 198775713.
  273. ^ Buruk, Banu; Ekmekci, Perihan Elif; Arda, Berna (1 September 2020). "A critical perspective on guidelines for responsible and trustworthy artificial intelligence". Medicine, Health Care and Philosophy. 23 (3): 387–399. doi:10.1007/s11019-020-09948-1. ISSN 1572-8633. PMID 32236794. S2CID 214766800.
  274. ^ Kamila, Manoj Kumar; Jasrotia, Sahil Singh (1 January 2023). "Ethical issues in the development of artificial intelligence: recognizing the risks". International Journal of Ethics and Systems. ahead-of-print (ahead-of-print). doi:10.1108/IJOES-05-2023-0107. ISSN 2514-9369. S2CID 259614124.
  275. ^ "AI Safety Institute releases new AI safety evaluations platform". UK Government. 10 May 2024. Retrieved 14 May 2024.
  276. ^ Regulation of AI to mitigate risks:
  277. ^ Jump up to: a b Vincent (2023).
  278. ^ Stanford University (2023).
  279. ^ Jump up to: a b c d UNESCO (2021).
  280. ^ Kissinger (2021).
  281. ^ Altman, Brockman & Sutskever (2023).
  282. ^ VOA News (25 October 2023). "UN Announces Advisory Body on Artificial Intelligence".
  283. ^ Edwards (2023).
  284. ^ Kasperowicz (2023).
  285. ^ Fox News (2023).
  286. ^ Milmo, Dan (3 November 2023). "Hope or Horror? The great AI debate dividing its pioneers". The Guardian Weekly. pp. 10–12.
  287. ^ "The Bletchley Declaration by Countries Attending the AI Safety Summit, 1–2 November 2023". GOV.UK. 1 November 2023. Archived from the original on 1 November 2023. Retrieved 2 November 2023.
  288. ^ "Countries agree to safe and responsible development of frontier AI in landmark Bletchley Declaration". GOV.UK (Press release). Archived from the original on 1 November 2023. Retrieved 1 November 2023.
  289. ^ "Second global AI summit secures safety commitments from companies". Reuters. 21 May 2024. Retrieved 23 May 2024.
  290. ^ "Frontier AI Safety Commitments, AI Seoul Summit 2024". gov.uk. 21 May 2024. Archived from the original on 23 May 2024. Retrieved 23 May 2024.
  291. ^ Jump up to: a b Russell & Norvig 2021, p. 9.
  292. ^ «Google книги ngram» .
  293. ^ Непосредственные предшественники ИИ:
  294. ^ Jump up to: а б Оригинальная публикация Тьюринга теста Тьюринга в « Вычислительной технике и интеллекте »: Историческое влияние и философские последствия:
  295. ^ Кревье (1993) , стр. 47–49.
  296. ^ Рассел и Норвиг (2003) , с. 17.
  297. ^ Рассел и Норвиг (2003) , с. 18.
  298. ^ Ньюквист (1994) , стр. 86–86.
  299. ^ Саймон (1965 , стр. 96), цитируется по Crevier (1993 , стр. 109).
  300. ^ Минский (1967 , стр. 2), цитируется по Crevier (1993 , стр. 109).
  301. ^ Рассел и Норвиг (2021) , с. 21.
  302. ^ Лайтхилл (1973) .
  303. ^ NRC 1999 , стр. 212–213.
  304. ^ Рассел и Норвиг (2021) , с. 22.
  305. ^ Экспертные системы :
  306. ^ Рассел и Норвиг (2021) , с. 24.
  307. ^ Нильссон (1998) , стр. 7.
  308. ^ МакКордак (2004) , стр. 454–462.
  309. ^ Моравец (1988) .
  310. ^ Jump up to: а б Брукс (1990) .
  311. ^ Развивающая робототехника :
  312. ^ Рассел и Норвиг (2021) , с. 25.
  313. ^
  314. ^ Рассел и Норвиг (2021) , с. 26.
  315. ^ Формальные и узкие методы, принятые в 1990-е годы:
  316. ^ ИИ широко использовался в конце 1990-х годов:
  317. ^ Люди (2023) .
  318. ^ Закон Мура и ИИ:
  319. ^ Jump up to: а б с Кларк (2015b) .
  320. ^ Большие данные :
  321. ^ Сагар, Рам (3 июня 2020 г.). «OpenAI выпускает GPT-3, самую большую модель на данный момент» . Журнал Analytics India . Архивировано из оригинала 4 августа 2020 года . Проверено 15 марта 2023 г.
  322. ^ ДиФелициантонио (2023) .
  323. ^ Госвами (2023) .
  324. ^ Jump up to: а б Тьюринг (1950) , с. 1.
  325. ^ Тьюринг (1950) , В разделе «Аргумент от сознания».
  326. ^ Рассел и Норвиг (2021) , с. 3.
  327. ^ Создатель (2006) .
  328. ^ Маккарти (1999) .
  329. ^ Минский (1986) .
  330. ^ «Что такое искусственный интеллект (ИИ)?» . Облачная платформа Google . Архивировано из оригинала 31 июля 2023 года . Проверено 16 октября 2023 г.
  331. ^ «Одна из самых больших проблем в регулировании ИИ – это согласование определения» . Carnegieendowment.org . Проверено 31 июля 2024 г.
  332. ^ «ИИ или чушь? Как определить, действительно ли маркетинговый инструмент использует искусственный интеллект» . Барабан . Проверено 31 июля 2024 г.
  333. ^ Нильссон (1983) , стр. 10.
  334. ^ Хаугеланд (1985) , стр. 112–117.
  335. ^ Гипотеза системы физических символов: Историческое значение:
  336. ^ Парадокс Моравеца :
  337. ^ Критика ИИ Дрейфусом : Историческое значение и философские последствия:
  338. ^ Кревье (1993) , с. 125.
  339. ^ Лэнгли (2011) .
  340. ^ Кац (2012) .
  341. ^ Аккуратность против неряшливости , исторические дебаты: Классический пример «неряшливого» подхода к разведке: Современный пример аккуратного ИИ и его стремлений в 21 веке:
  342. ^ Пенначин и Герцель (2007) .
  343. ^ Jump up to: а б Робертс (2016) .
  344. ^ Рассел и Норвиг (2021) , с. 986.
  345. ^ Чалмерс (1995) .
  346. ^ Деннетт (1991) .
  347. ^ Хорст (2005) .
  348. ^ Сирл (1999) .
  349. ^ Сирл (1980) , с. 1.
  350. ^ Рассел и Норвиг (2021) , с. 9817.
  351. ^ Аргумент Сирла в китайской комнате : Обсуждение:
  352. ^ Лейт, Сэм (7 июля 2022 г.). «Ник Бостром: Как мы можем быть уверены, что машина не находится в сознании?» . Зритель . Проверено 23 февраля 2024 г.
  353. ^ Jump up to: а б с Томсон, Джонни (31 октября 2022 г.). «Почему у роботов нет прав?» . Большое Думай . Проверено 23 февраля 2024 г.
  354. ^ Jump up to: а б Кейтман, Брайан (24 июля 2023 г.). «ИИ должен бояться людей» . Время . Проверено 23 февраля 2024 г.
  355. ^ Вонг, Джефф (10 июля 2023 г.). «Что лидерам нужно знать о правах роботов» . Компания Фаст .
  356. ^ Херн, Алекс (12 января 2017 г.). «Придать роботам статус личности, - утверждает комитет ЕС» . Хранитель . ISSN   0261-3077 . Проверено 23 февраля 2024 г.
  357. ^ Дови, Дана (14 апреля 2018 г.). «Эксперты не считают, что у роботов должны быть права» . Newsweek . Проверено 23 февраля 2024 г.
  358. ^ Кадди, Элис (13 апреля 2018 г.). «Права роботов нарушают права человека, предупреждают эксперты ЕС» . Евроньюс . Проверено 23 февраля 2024 г.
  359. ^ Интеллектуальный взрыв и технологическая сингулярность : Эй Джей Гуда «Интеллектуальный взрыв» Вернора Винджа «Необычность»
  360. ^ Рассел и Норвиг (2021) , с. 1005.
  361. ^ Трансгуманизм :
  362. ^ ИИ как эволюция:
  363. ^ ИИ в мифе:
  364. ^ МакКордак (2004) , стр. 340–400.
  365. ^ Бутаццо (2001) .
  366. ^ Андерсон (2008) .
  367. ^ МакКоли (2007) .
  368. ^ Гальван (1997) .

Учебники по искусственному интеллекту

Два наиболее широко используемых учебника в 2023 году (см. Открытую программу ):

Это были четыре наиболее широко используемых учебника по ИИ в 2008 году:

Более поздние издания:

История ИИ

Другие источники

Дальнейшее чтение

Arc.Ask3.Ru: конец переведенного документа.
Arc.Ask3.Ru
Номер скриншота №: 4fe780c1c12c1bf697b5bdc4da09d7c4__1723183080
URL1:https://arc.ask3.ru/arc/aa/4f/c4/4fe780c1c12c1bf697b5bdc4da09d7c4.html
Заголовок, (Title) документа по адресу, URL1:
Artificial intelligence - Wikipedia
Данный printscreen веб страницы (снимок веб страницы, скриншот веб страницы), визуально-программная копия документа расположенного по адресу URL1 и сохраненная в файл, имеет: квалифицированную, усовершенствованную (подтверждены: метки времени, валидность сертификата), открепленную ЭЦП (приложена к данному файлу), что может быть использовано для подтверждения содержания и факта существования документа в этот момент времени. Права на данный скриншот принадлежат администрации Ask3.ru, использование в качестве доказательства только с письменного разрешения правообладателя скриншота. Администрация Ask3.ru не несет ответственности за информацию размещенную на данном скриншоте. Права на прочие зарегистрированные элементы любого права, изображенные на снимках принадлежат их владельцам. Качество перевода предоставляется как есть. Любые претензии, иски не могут быть предъявлены. Если вы не согласны с любым пунктом перечисленным выше, вы не можете использовать данный сайт и информация размещенную на нем (сайте/странице), немедленно покиньте данный сайт. В случае нарушения любого пункта перечисленного выше, штраф 55! (Пятьдесят пять факториал, Денежную единицу (имеющую самостоятельную стоимость) можете выбрать самостоятельно, выплаичвается товарами в течение 7 дней с момента нарушения.)