The phrase "artificial intelligence" was put forward as a field name in the title of a research proposal dated August 31, 1955, and every other name used for the field since then points at a different part of it: cybernetics for control and feedback, machine learning for programs that improve from data, augmented intellect for tools that extend human thinking, and foundation models for large systems trained once and adapted to many tasks. These terms are not interchangeable synonyms. Each one names a narrower idea, a different goal, or a newer way of building systems.
Key takeaways
- "Artificial intelligence" first appeared in a 1955 funding proposal for a summer workshop at Dartmouth, written by four researchers.
- "Cybernetics" (Norbert Wiener, 1948) and "machine learning" (Arthur Samuel, 1959) describe parts of the field, not the whole.
- "Foundation model" was introduced in a 2021 Stanford report and refers to a training approach, not a separate kind of intelligence.
- A product that swaps the label "AI" for a softer term has not changed what it does, so the question to ask is what the system actually computes.
Where the name "artificial intelligence" came from
The proposal for the Dartmouth Summer Research Project on Artificial Intelligence, is dated August 31, 1955, and was reprinted in AI Magazine in 2006. Its authors were McCarthy, Minsky, Rochester, and Shannon. The plan was a two-month study by 10 people in the summer of 1956, resting on the conjecture that learning and other features of intelligence could be described exactly enough for a machine to imitate them. Its list of topics included automatic computers, neuron nets, self-improvement, abstractions, and randomness and creativity.
The proposal shows that the name was a program of research and not a product category. Accounts of why the authors chose the words differ among secondary sources, so this page does not attribute a motive to them.
Older and parallel names
Several terms were in use before or alongside the Dartmouth name, and some remain in use. Alan Turing wrote a 1948 report for Britain's National Physical Laboratory titled "Intelligent Machinery" and a 1950 paper in the journal Mind titled "Computing Machinery and Intelligence." Norbert Wiener published "Cybernetics: Or Control and Communication in the Animal and the Machine" in 1948, treating control and communication in animals and machines as one subject. Arthur Samuel's 1959 paper in the IBM Journal of Research and Development, "Some Studies in Machine Learning Using the Game of Checkers," reported a program that could learn to play better checkers than its author. Douglas Engelbart's 1962 report "Augmenting Human Intellect," prepared at SRI, framed computers as ways to increase a person's capability to approach a complex problem.
A modern name: foundation models
In August 2021, a group of 114 authors led by Rishi Bommasani posted "On the Opportunities and Risks of Foundation Models" on arXiv. The report coined the term for models trained on broad data that can be adapted to many downstream tasks, and the authors wrote that the name underscores the models' "critically central yet incomplete character." Large language models are one example. The term describes how a system is built and reused, not whether it is intelligent.
Comparison of the main terms
| Term | Early source | What it points at | Not the same as |
|---|---|---|---|
| Artificial intelligence | Dartmouth proposal, 1955 | Machines simulating features of intelligence | Any one technique |
| Cybernetics | Wiener, 1948 | Control and communication in animals and machines | Today's neural networks |
| Machine learning | Samuel, 1959 | Programs that improve their play or predictions from experience | All AI, since rule-based programs exist |
| Augmenting human intellect | Engelbart, 1962 | Tools that extend what a person can do | Machines acting on their own |
| Foundation model | Bommasani et al., 2021 | Broadly trained models adapted to many tasks | A claim about general intelligence |
Three checks before trusting a new label
- Does the description say what data the system learned from, or what rules it follows?
- Is the term a goal (simulate intelligence), a method (learn from data), or a way of use (assist a person)?
- Does the same product page describe an outcome that can be tested, such as accuracy on a stated task?
An illustrative example: a vendor describes its tool as "cognitive computing" instead of AI. The new phrase gives no information. The useful question is what it does with inputs, which only the technical description can answer.
FAQ
Is machine learning the same as artificial intelligence?
No. Machine learning is one way to build AI systems, in which a program improves from data. Rule-based programs can also be called AI in the broad sense, and the name "machine learning" dates to Samuel's 1959 paper.
Is cybernetics another word for AI?
No. Wiener's 1948 book treated control and communication in animals and machines as one subject. That overlaps with AI in places, but it is a different field with a different center of attention.
Is a foundation model the same as a chatbot?
No. A foundation model is a broadly trained base. A chatbot is an application that may be built on one, with extra training and an interface added.
Which term is the most accurate?
None is accurate for everything. Terms are accurate when they match the thing described: "machine learning" for systems trained on data, "foundation model" for broadly trained base models, and "AI" as the umbrella.
Related reading
- Alternatives to artificial intelligence
- What vectors do inside AI systems
- The components an AI system needs
- A university guide to what AI is
More in AI Basics.
Last verified: October 2026
Sources
- McCarthy J, Minsky ML, Rochester N, Shannon CE, A Proposal for the Dartmouth Summer Research Project on Artificial Intelligence, August 31, 1955; AI Magazine 27(4), 2006
- Turing AM, Intelligent Machinery, National Physical Laboratory report, 1948
- Turing AM, Computing Machinery and Intelligence, Mind 59(236), 1950
- Wiener N, Cybernetics: Or Control and Communication in the Animal and the Machine, 1948
- Samuel AL, Some Studies in Machine Learning Using the Game of Checkers, IBM Journal of Research and Development 3(3), 1959
- Engelbart DC, Augmenting Human Intellect: A Conceptual Framework, SRI, 1962
- Bommasani R et al., On the Opportunities and Risks of Foundation Models, arXiv:2108.07258, 2021