Beginner · 11
A short history of AI
From a thought experiment in 1950 to chatbots in your pocket: the big moments, the long winters and why it all sped up.
6 min read · Tap underlined words for a definition
Swipe up to begin
AI feels brand new, but people have been working on it for over seventy years. The story is not a straight line. It is a series of big hopes, long disappointments and a few sudden leaps.
Where it began
In 1950 the mathematician Alan Turing asked a simple question: can machines think? He suggested a game instead of a definition. If a person chatting by text can't tell a machine from a human, perhaps that is good enough. We now call this the Turing test.
In 1956 a small group of researchers met for a summer workshop at Dartmouth College in the USA. The name "" comes from their proposal, and many people treat that summer as the birth of the field.
Here are the big moments, in order:
- 1950: Turing publishes "Computing Machinery and Intelligence".
- 1956: The Dartmouth workshop gives the field its name.
- 1970s: Early promises go unmet and funding dries up: the first AI winter.
- 1980s: Expert systems boom: programs built from hand-written rules by human experts.
- Late 1980s to early 1990s: prove costly and brittle. A second winter follows.
- 1997: IBM's Deep Blue beats world chess champion Garry Kasparov.
- 2012: A called AlexNet wins the ImageNet image recognition contest by a wide margin. takes off.
- 2016: DeepMind's AlphaGo beats Lee Sedol, one of the world's best Go players.
- 2017: Google researchers publish the transformer, the design behind today's .
- 2022: OpenAI releases ChatGPT, and AI chat reaches the public almost overnight.
What was an "AI winter"?
Rules versus learning
Expert systems tried to write intelligence down as rules: "if the patient has a fever and a rash, consider measles". This worked in narrow areas, but the real world has too many exceptions to list.
Modern AI flips the approach. Instead of writing rules, we show a system lots of examples and let it learn the patterns itself.
Why it sped up after 2012
The ideas behind neural networks are decades old. What changed was three things arriving together: huge amounts of data from the internet, powerful graphics chips that could do the maths quickly, and better training methods.
Which event in 2012 is often seen as the start of the deep learning boom?
Before you drift off
- AI began as a research field in the 1950s, with Turing and the Dartmouth workshop.
- Hand-written rules hit their limits, leading to two AI winters.
- Data, fast chips and better methods made learning from examples work, from AlexNet to ChatGPT.
Terms in this lesson
Go deeper, free and open
- History of artificial intelligence · Wikipedia
- The Turing Test · Stanford Encyclopedia of Philosophy
- Attention Is All You Need · Vaswani et al., 2017 · The transformer paper
That's the lesson
A short history of AI · 6 min
Spotted a mistake? Suggest an edit
Previous: Using AI wisely: privacy, bias, judgement