Who Invented AI?

The Origins of Artificial Intelligence

Artificial intelligence was not invented by a single individual in a single moment. Instead, it emerged as an academic discipline through the collective work of mathematicians, logicians, neurophysiologists, and computer scientists during the mid-twentieth century. However, when historians and researchers point to the formal inception of the field and the coining of the term "artificial intelligence," American computer scientist and mathematician John McCarthy is widely recognized as the primary organizer who named and founded AI as a distinct discipline. Dartmouth workshop - Wikipedia John McCarthy (computer scientist)

In 1955, McCarthy, then a mathematics professor at Dartmouth College, authored a proposal for a summer research project alongside Marvin Minsky, Nathaniel Rochester, and Claude Shannon. Held at Dartmouth in the summer of 1956, the Dartmouth Summer Research Project on Artificial Intelligence became the historic milestone where AI was organized into an autonomous field of scientific inquiry. Dartmouth workshop - Wikipedia The Birth of “Artificial Intelligence”

While McCarthy coined the phrase and organized the field, several other key pioneers created the theoretical foundations, mathematical frameworks, and functional computer programs that made artificial intelligence possible. John McCarthy (computer scientist)

Key Figures in the Creation of AI

Understanding who invented AI requires examining the foundational contributors across three critical phases: theoretical foundations, early computational models, and the birth of symbolic AI programs.

+-------------------------------------------------------------------------+
|                        PIONEERS OF ARTIFICIAL INTELLIGENCE              |
+------------------------------------+------------------------------------+
| Pioneer                            | Primary Contribution               |
+------------------------------------+------------------------------------+
| Alan Turing                        | Computational theory & Turing Test |
| Warren McCulloch & Walter Pitts    | First mathematical neural network  |
| John McCarthy                      | Coined "AI" & organized Dartmouth  |
| Marvin Minsky                      | Neural networks & frame theory     |
| Allen Newell, Herbert Simon,       | Logic Theorist (first running AI)  |
|   & Cliff Shaw                     |                                    |
| Claude Shannon                     | Information theory & game theory   |
+------------------------------------+------------------------------------+

Alan Turing: Theoretical Foundations and Machine Intelligence

British mathematician and logician Alan Turing laid the intellectual groundwork for computer science and artificial intelligence well before electronic digital computers were widely available. In his 1936 paper introducing the theoretical "Turing machine," Turing proved that a universal computing device could execute any mathematical algorithm that could be formally defined. The birth of Artificial Intelligence (AI) research - LLNL

In 1950, Turing published his seminal paper "Computing Machinery and Intelligence", opening with the question: "Can machines think?" To evaluate whether a machine possesses intelligence without getting bogged down in defining "thinking," he proposed the "Imitation Game," now known as the Turing Test. Turing argued that if a human interrogator cannot reliably distinguish machine-generated responses from human responses through text-based dialogue, the machine exhibits intelligent behavior. The birth of Artificial Intelligence (AI) research - LLNL

Warren McCulloch and Walter Pitts: The First Artificial Neural Network

In 1943, neurophysiologist Warren McCulloch and logician Walter Pitts published "A Logical Calculus of the Ideas Immanent in Nervous Activity". This work provided the first mathematical abstraction of biological neurons. Birthplace of Neural Networks: McCulloch & Pitts at UChicago

McCulloch and Pitts demonstrated that simplified networks of binary threshold units—neurons that either fire or remain dormant based on their inputs—could perform fundamental logical operations such as AND, OR, and NOT. This model established the connectionist approach to AI and proved mathematically that neural structures could carry out logical computations. Birthplace of Neural Networks: McCulloch & Pitts at UChicago

The Dartmouth Workshop (1956): Coining and Founding AI

In August 1955, John McCarthy, Marvin Minsky (then at Harvard), Nathaniel Rochester (IBM), and Claude Shannon (Bell Telephone Laboratories) submitted a formal grant proposal to the Rockefeller Foundation. The proposal contained the foundational premise of the entire field:

"The study is to proceed on the basis of the conjecture that every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it." A Proposal for the Dartmouth Summer Research Project on Artificial ...

The resulting workshop took place during the summer of 1956 at Dartmouth College in Hanover, New Hampshire. It brought together ten core researchers to examine language processing, neural networks, computational complexity, abstraction, and self-improvement. The Dartmouth conference separated AI from existing umbrella fields such as cybernetics, automata theory, and operations research, establishing it as an independent branch of computer science. Dartmouth workshop - Wikipedia Cybernetics, Automata Studies, and the Dartmouth Conference on ...

Allen Newell, Herbert A. Simon, and Cliff Shaw: The First Working AI Program

While the Dartmouth conference established the name and research agenda, the first actual computer program demonstrating artificial intelligence arrived courtesy of Allen Newell, Herbert A. Simon, and Cliff Shaw. The first AI ever built and why Logic Theorist still matters

In 1955 and 1956, the trio developed the Logic Theorist at the RAND Corporation and Carnegie Institute of Technology (now Carnegie Mellon University). Designed to mimic human problem-solving heuristics, Logic Theorist searched through trees of mathematical possibilities to prove theorems from Alfred North Whitehead and Bertrand Russell’s Principia Mathematica. It successfully proved 38 of the first 52 theorems in chapter two of the work, often finding shorter and more elegant proofs than the original authors. Because it manipulated symbolic representations rather than merely calculating numbers, Logic Theorist is widely cited as the first true artificial intelligence software. Cybernetics, Automata Studies, and the Dartmouth Conference on ... The first AI ever built and why Logic Theorist still matters

Evolution of AI Paradigms

The emergence of AI spawned two primary philosophical and architectural traditions, both tracing back to the original inventors.

Historical Timeline of AI Inception

The development of artificial intelligence followed a chronological sequence of theoretical breakthroughs and practical implementations:

1. 1936: Alan Turing introduces the universal Turing machine, establishing computability limits. The birth of Artificial Intelligence (AI) research - LLNL
2. 1943: Warren McCulloch and Walter Pitts publish the first mathematical formulation of an artificial neural network. Birthplace of Neural Networks: McCulloch & Pitts at UChicago
3. 1950: Alan Turing publishes "Computing Machinery and Intelligence" and introduces the Turing Test. The birth of Artificial Intelligence (AI) research - LLNL
4. 1951: Marvin Minsky and Dean Edmonds construct SNARC (Stochastic Neural Analog Reinforcement Calculator), the first neural network hardware machine. Dartmouth workshop - Wikipedia
5. 1955: Allen Newell, Herbert A. Simon, and Cliff Shaw complete the Logic Theorist program. The first AI ever built and why Logic Theorist still matters
6. 1956: John McCarthy organizes the Dartmouth Summer Research Project, formally establishing artificial intelligence as an academic discipline. Dartmouth workshop - Wikipedia The Birth of “Artificial Intelligence”
7. 1958: McCarthy invents Lisp, the primary programming language for AI research for several decades, and develops the concept of time-sharing systems. John McCarthy (computer scientist)

Summary of Historical Attribution

Attributing the invention of artificial intelligence depends on how "invention" is defined:

• Conceptual Visionary: Alan Turing invented the computational theory of machine intelligence and formulated how to test it. The birth of Artificial Intelligence (AI) research - LLNL
• Discipline Founder: John McCarthy coined the term "Artificial Intelligence" and organized the community that launched the field. Dartmouth workshop - Wikipedia John McCarthy (computer scientist)
• First Software Implementers: Allen Newell, Herbert A. Simon, and Cliff Shaw built the Logic Theorist, proving that computers could perform symbolic reasoning beyond arithmetic. The first AI ever built and why Logic Theorist still matters
• Mathematical Modeler: Warren McCulloch and Walter Pitts created the first synthetic architecture modeling neural logic. Birthplace of Neural Networks: McCulloch & Pitts at UChicago

Artificial intelligence is therefore a collective scientific invention, catalyzed by Turing's mid-century theory, codified by McCarthy at Dartmouth in 1956, and operationalized by the early pioneers of symbolic and neural computation. Dartmouth workshop - Wikipedia John McCarthy (computer scientist) The birth of Artificial Intelligence (AI) research - LLNL

The short answer: AI was named by John McCarthy, but not invented by any one person

If the question is who coined the term "artificial intelligence," the answer is precise: the American computer scientist John McCarthy, who used the phrase in an August 31, 1955 proposal for a summer research project held at Dartmouth College the following year. That proposal, co-authored with Marvin Minsky, Nathaniel Rochester, and Claude Shannon, gave the field its name and its founding agenda. Artificial Intelligence (AI) Coined at Dartmouth [PDF] A Proposal for the Dartmouth Summer Research Project on Artificial ... Dartmouth workshop - Wikipedia

If the question is who invented the underlying ideas, there is no single inventor. Artificial intelligence is a research field, not a device with a patent. Its core ingredients — formal logic, the theory of computation, the idea that a machine could carry out reasoning, the mathematical model of a neuron, learning from data — were contributed by different people across several decades and disciplines. The most defensible answer is that AI has founders rather than an inventor, and that the leading candidates for "who invented AI first" are Alan Turing (who framed the question and its test), Warren McCulloch and Walter Pitts (who built the first mathematical model of neural computation), John McCarthy (who named and organized the field), and Allen Newell, Herbert Simon and Cliff Shaw (who wrote the first working AI program).

Why "invented" is a slippery word here

Attribution disputes in AI usually come from mixing up four different questions that people ask with the same words:

Question	Best answer
Who coined the phrase "artificial intelligence"?	John McCarthy, in the 1955 Dartmouth proposal Artificial Intelligence (AI) Coined at Dartmouth
Who first argued that machines could think, and how we'd know?	Alan Turing, in his 1950 paper "Computing Machinery and Intelligence" COMPUTING MACHINERY AND INTELLIGENCE The birth of Artificial Intelligence (AI) research - LLNL
Who created the first artificial neural network model?	Warren McCulloch and Walter Pitts, 1943 A Logical Calculus of the Ideas Immanent in Nervous Activity Birthplace of Neural Networks: McCulloch & Pitts at UChicago
Who wrote the first AI program?	Allen Newell, Herbert Simon and Cliff Shaw's Logic Theorist, completed in 1956 Logic Theorist - Wikipedia History - The Simon Initiative - Carnegie Mellon University

Different institutions and textbooks emphasize different milestones, so the label "father of AI" is applied variously to Turing, McCarthy, and Minsky depending on which of these questions the writer has in mind. That inconsistency is not sloppiness so much as a genuine feature of how the field emerged: the concept, the mathematics, the name, and the first machine demonstrations arrived from separate directions within about fifteen years.

The long prehistory: mechanizing thought

The ambition behind AI is far older than the computer. Aristotle's syllogistic logic set out rules by which valid conclusions follow from premises without appeal to intuition — the first serious suggestion that reasoning has a formal structure. In the thirteenth century Ramon Llull built rotating paper discs intended to generate combinations of concepts mechanically. Leibniz dreamed of a calculus ratiocinator, a symbolic language in which disputes could be settled by calculation. George Boole's nineteenth-century algebra of logic turned that dream into workable mathematics, and Charles Babbage's designs for the Analytical Engine, together with Ada Lovelace's notes on it, established that a machine could in principle manipulate symbols of any kind, not just numbers. Lovelace also raised the first great skeptical objection: that a machine can only do what it is ordered to do — an objection Turing would later answer directly.

The decisive twentieth-century step was the theory of computation. Turing's 1936 analysis of computable numbers defined an abstract machine capable of carrying out any effective procedure, which implied that a single general-purpose device could, given the right program, simulate any well-specified process — including, perhaps, processes we call thinking. Once that idea was joined to working electronic computers in the 1940s, the question stopped being philosophical and became an engineering programme.

Turing's framing: "Can machines think?"

Alan Turing is often called the father of AI because he was the first to treat machine intelligence as a serious, testable research problem. His 1950 paper Computing Machinery and Intelligence, published in the journal Mind, opens by proposing to consider the question "Can machines think?" and then replaces it with an operational substitute he called the imitation game: an interrogator communicates by text with a human and a machine and tries to tell which is which. The procedure is now universally known as the Turing test. COMPUTING MACHINERY AND INTELLIGENCE Turing test The birth of Artificial Intelligence (AI) research - LLNL

What makes the paper foundational is not the test alone but the surrounding discussion. Turing anticipated and answered objections from theology, mathematics, consciousness, and Lovelace's "machines cannot originate anything," and he sketched the idea of a child machine that would be educated rather than fully programmed — an early statement of the case for machine learning. He also worked on the topic earlier, in an unpublished internal report on intelligent machinery written while at the National Physical Laboratory in the late 1940s, and he thought about computer chess before hardware existed to run it. Turing died in 1954, two years before the Dartmouth meeting, so he never participated in the field he had done so much to define. Turing test

McCulloch and Pitts: the first artificial neurons

Independently of the logic tradition, the neurophysiologist Warren McCulloch and the logician Walter Pitts published A Logical Calculus of the Ideas Immanent in Nervous Activity in 1943, modelling the nervous system as a network of simple all-or-none logical elements. These McCulloch–Pitts neurons were the first mathematical model of a neural network, and they showed that networks of such elements could compute logical functions. A Logical Calculus of the Ideas Immanent in Nervous Activity A logical calculus of the ideas immanent in nervous activity Birthplace of Neural Networks: McCulloch & Pitts at UChicago

This paper is the ancestor of everything now called deep learning, and it is why "who invented AI" sometimes gets the answer "McCulloch and Pitts." Their model had no learning rule, however; it described what a network could compute, not how it could acquire that ability. Frank Rosenblatt's perceptron, introduced in 1958, added a training procedure that adjusted connection weights from examples, making it the first widely known trainable neural network for pattern recognition. Rosenblatt - an overview | ScienceDirect Topics

Dartmouth 1956: the field gets a name

In the summer of 1955 McCarthy, then at Dartmouth, sought funding from the Rockefeller Foundation for a two-month, ten-man study. The resulting proposal, dated August 31, 1955 and signed by McCarthy, Minsky (then at Harvard), Rochester (IBM) and Shannon (Bell Telephone Laboratories), contains the first use of the term in this technical sense and states the working assumption of the whole enterprise: that every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it. Artificial Intelligence (AI) Coined at Dartmouth [PDF] A Proposal for the Dartmouth Summer Research Project on Artificial ... Dartmouth workshop - Wikipedia

McCarthy is generally understood to have chosen the phrase partly to mark out a distinct identity from cybernetics and automata theory, which were then the dominant labels for machine-and-mind research. The 1956 workshop itself was not a tidy conference: participants came and went over the weeks, there was no consensus outcome, and no shared methodology emerged. Its lasting importance is sociological — it assembled the people who would run the major American AI laboratories for the next thirty years and gave them a common banner. Dartmouth workshop - Wikipedia The Birth of “Artificial Intelligence”

The most concrete result presented there came from outside the organizing group. Allen Newell, Herbert Simon and Cliff Shaw of Carnegie Institute of Technology and RAND brought the Logic Theorist, a program deliberately engineered to perform automated reasoning; it proved theorems from Principia Mathematica and is widely described as the first artificial intelligence program. Newell and Simon first described it in a RAND report issued in June 1956. Notably, Newell and Simon preferred their own term, complex information processing, and were not enthusiastic about McCarthy's label — a reminder that the name was contested from the beginning. Logic Theorist - Wikipedia History - The Simon Initiative - Carnegie Mellon University Newell, Simon & Shaw Develop the First Artificial Intelligence Program

The founding generation and what each contributed

• John McCarthy — named the field, organized Dartmouth, invented the Lisp programming language (the dominant AI language for decades), contributed to time-sharing, and founded the AI laboratories at MIT (with Minsky) and later Stanford. He received the ACM Turing Award in 1971. John McCarthy (computer scientist)
• Marvin Minsky — co-proposer of Dartmouth, co-founder of the MIT AI Lab, builder of an early neural-network learning machine as a graduate student, and later a major critic of perceptrons and theorist of knowledge representation.
• Allen Newell and Herbert Simon — created the first reasoning programs and the physical symbol system view of intelligence; both were Turing Award recipients, and Simon also won the Nobel Memorial Prize in Economics.
• Claude Shannon — founder of information theory, builder of a maze-solving mechanical mouse, and author of early work on computer chess. Dartmouth workshop - Wikipedia
• Arthur Samuel — at IBM, wrote a checkers program that improved through self-play and popularized the term machine learning.
• Frank Rosenblatt — the perceptron and the connectionist line of research. Rosenblatt - an overview | ScienceDirect Topics
• Oliver Selfridge, Nathaniel Rochester, Ray Solomonoff, Trenchard More — Dartmouth participants whose work in pattern recognition, machine design and inductive inference shaped early directions.

Who invented the AI people use today?

Modern systems — image recognizers, translation engines, chatbots — descend more from the neural-network branch than from the symbolic branch that dominated the 1960s and 1970s. That lineage runs from McCulloch and Pitts through Rosenblatt, through the backpropagation-trained multilayer networks popularized in the 1980s, to the deep learning era. Geoffrey Hinton, Yann LeCun and Yoshua Bengio shared the 2018 ACM Turing Award for their work on deep neural networks, and they are frequently described as the field's modern architects; Hinton and John Hopfield were later awarded the 2024 Nobel Prize in Physics for foundational work on machine learning with artificial neural networks.

Two further steps matter for today's systems. The success of deep convolutional networks on large image datasets around 2012 demonstrated that scale — more data, more compute — could unlock capabilities that hand-crafted methods could not. Then the transformer architecture, introduced by a group of Google researchers in 2017, became the basis for large language models, and consumer chat interfaces built on such models brought the technology to a mass audience from late 2022 onward. None of these were invented by one person either; each is a collaborative paper with multiple authors building on prior work.

So the honest answer to "who invented AI first" depends on the definition you accept. The name belongs to McCarthy; the founding question and its test belong to Turing; the first neural model to McCulloch and Pitts; the first program to Newell, Simon and Shaw; and the technology now in everyday use to a much larger community of researchers whose contributions resist reduction to a single inventor.