AI Glossary / AI foundations / Foundation

Dartmouth workshop

Definition

The Dartmouth workshop was a summer research project held at Dartmouth College in 1956 that is widely treated as the founding event of artificial intelligence as a named field. A small group of researchers gathered to study whether features of human intelligence could be described precisely enough for a machine to simulate them. It was a planned meeting of about two months, not a product, a theory or a single breakthrough.

What it means

The 1956 workshop matters mostly for two things it did: it gave the field its name, and it set an ambitious goal. The proposal that led to it used the phrase "artificial intelligence" and stated the working assumption that learning and other aspects of intelligence could in principle be described so exactly that a machine could be made to simulate them.

For a business reader, the workshop is useful as a starting point for the story, not as a source of tools. Almost nothing built at the meeting is in use today. What carried forward was a way of framing the question and a community of people who kept working on it. Accounts differ on how much the meeting itself produced compared with what it symbolized, so it is best described as a founding moment and not as a technical turning point.

How it works

A workshop is a format, not a mechanism, so the useful question is how this one was organized. A small number of organizers wrote a proposal and sought funding. They invited researchers with interests in areas such as computing machines, language, neural networks, the theory of computation and self-improving programs. Attendance was informal. People came and went over the weeks, and there was no single agreed program or shared set of results at the end.

The proposal listed topics the group wanted to explore. These included how to program computers to use language, how networks of simple units might form concepts, how a machine might improve itself, and how randomness and creativity might fit in. Each of these themes became a research direction that later groups pursued, often under different names.

The lasting effect came from people and vocabulary. Participants went on to lead early research labs and teaching programs. The name they used spread to universities, funders and the press. Early optimism about how fast progress would come later proved too high, which contributed to cycles of enthusiasm and disappointment that historians of the field describe.

Three practical examples

• Finance team: A controller preparing a short briefing on AI for the board opens with the history. She notes that the term dates to a 1956 workshop and that early predictions were optimistic, then uses that to frame why the team will test claims with small pilots before committing budget.
• Retailer: A merchandising manager reads vendor materials that call a recommendation tool "cutting-edge AI". Knowing the field's long history, from the 1956 workshop onward, he asks what the tool actually does and how it is measured, since the label has covered many different techniques over the decades.
• Support desk: A training lead builds a one-page timeline for new staff who will work with a chatbot. The 1956 Dartmouth workshop is the first entry, followed by later milestones, so staff see the technology as the product of many decades of work and not a sudden arrival.

Think of it this way

Think of the Dartmouth workshop as the christening of a child: the name was chosen, the guests gathered and hopes were stated, but the child's actual abilities developed over many years afterward. The analogy stops working because a christening is a single ceremony with one family, while the workshop was a loose gathering of researchers who did not share one plan or reach one agreed result.

Worked example

A training manager wants to show how long the field has existed. She uses 1956 as the start and a hypothetical current year of 2025.

Step 1: Subtract the start year from the current year. 2025 minus 1956 equals 69 years.

Step 2: Divide the span into chunks for a slide. If she marks every 23 years, 69 divided by 23 equals 3 chunks, so the timeline has three equal periods.

Step 3: She places three example milestones, one per period, and labels the workshop as the starting point at year 0.

The point for staff is simple: the field is roughly 69 years old in this example, so the idea of AI is old even though everyday tools are recent. The numbers here are illustrative and chosen only to make the arithmetic easy to follow.

Illustrative case study

Fictional case: Larchwood Logistics, a mid-sized regional freight broker, was asked by its owners to adopt AI for route planning. The operations director, Mei, began with a vendor who described the product as the realization of a goal set in 1956 and promised that most planning work would be automated within a year.

Mei had read a short history of the field and recognized the pattern of confident early forecasts. She asked the vendor for a limited trial on one depot's routes for eight weeks, with a clear measure: average delivery time compared with the existing schedule. The trial showed modest gains on simple routes and no gains on routes with frequent last-minute changes. The vendor's year-one promise had been far too high.

Instead of dropping the project, Larchwood narrowed the scope. It used the tool for simple routes, kept human planners for complex ones, and reviewed results quarterly. The lesson the team took from the history was to treat ambitious claims as hypotheses and to test them on small, measurable tasks.

Common misconceptions and mistakes

The Dartmouth workshop invented the first working AI system. It did not; it was a research gathering that named the field and set a goal, and it produced no deployable product.

The workshop was a large, formal conference with agreed conclusions. It was a small, informal summer project with changing attendance and no single shared result.

AI began suddenly with recent chatbots. The field has a history that goes back to at least 1956, and many ideas behind current tools are decades old.

The founders correctly predicted how fast progress would come. Early forecasts were often too optimistic, and that gap between promise and result is a recurring theme in the field's history.

Business use

The practical use of this entry is as context. When you evaluate AI claims, remembering that the field has seen waves of optimism helps you ask for evidence, define a narrow pilot and set a measurable target. A short timeline that starts in 1956 can also help onboard staff and calm both excessive hype and excessive fear.

The cost is close to zero: a little reading time. The risk is over-reading history, for example assuming that past disappointments mean current tools will disappoint, or treating the workshop's founding goal as a guarantee. Measure what a specific tool does on your own data and tasks, and treat the history as background and not as a forecast.

People also ask

Why is the Dartmouth workshop important?

It is widely credited with giving artificial intelligence its name and with stating the goal of describing intelligence precisely enough for a machine to simulate it. Its importance is mainly historical and symbolic.

When was the Dartmouth workshop held?

It took place in the summer of 1956 at Dartmouth College and ran for roughly two months. Attendance was informal, with people joining for different parts of the period.

Does the Dartmouth workshop affect AI tools I use today?

Not directly, since no current product comes from it. It matters because it framed the field's goals and helped form the research community that later built the techniques behind modern tools.

Check your understanding

What two things is the Dartmouth workshop mainly remembered for?

Answer: It gave the field the name artificial intelligence and set the ambitious goal of describing intelligence so precisely that a machine could simulate it.

In the worked example, how many years separate 1956 and 2025?

Answer: 69 years, found by subtracting 1956 from 2025.

Why did Larchwood Logistics run an eight-week trial?

Answer: The director saw the vendor's year-one promise as a hypothesis to test, so she used a small, measurable trial on one depot before committing.

Did the workshop produce a deployable AI product?

Answer: No. It was a small research gathering, and its lasting effect came from the name, the goal and the community it helped form.

Related terms

• Artificial intelligence
• AI winter
• Symbolic AI
• Machine learning
• Neural network
• Turing test
• Expert system
• Generative AI

Learn next

Try writing a one-page timeline of AI for your own team, marking the 1956 workshop as the start and adding two or three later milestones you find in the sources below. Then read the entries on AI winter and symbolic AI to see how early optimism and later setbacks shaped the field.

More in AI foundations

• Code model - A code model is an AI model trained or tuned mainly on source code and related text so that it can read, write, explain and fix programs.
• Common Crawl - Common Crawl is a nonprofit organization and the large, openly available archive of web pages it collects by crawling the public internet and releases for...
• Context window - A context window is the limited amount of information a language model can work with in a particular generation.
• Data annotation - Data annotation is the process of adding labels, tags or notes to raw data such as text, images, audio, video or records, so that a machine learning system...
• Data labeling - Data labeling is the process of attaching tags, categories, scores or other annotations to raw data such as text, images, audio or records, so that a...
• Deep learning - Deep learning is a branch of machine learning that learns representations through multiple layers of computation.

Sources

Russell and Norvig, Artificial Intelligence: A Modern Approach: https://aima.cs.berkeley.edu/ Stanford HAI, AI Index: https://hai.stanford.edu/ai-index ISO, ISO/IEC 22989 Artificial intelligence concepts and terminology: https://www.iso.org/standard/74296.html The sources were consulted for background on the history and terminology of artificial intelligence; the examples, analogy, worked example and case are original educational illustrations. Checked 5 Nov 2026.

Previous: Dark pattern Next: Data access