What is the history of artificial intelligence?

Artificial intelligence can feel brand new, but its story stretches back over 70 years. Understanding that history makes today’s breakthroughs — and the hype around them — much easier to read.

Where did the idea of AI begin?

The modern story starts with mathematician Alan Turing. In 1950 he asked a deceptively simple question — “Can machines think?” — and proposed a practical test for machine intelligence, now known as the Turing test.

A few years later, in 1956, a summer workshop at Dartmouth College brought together researchers who believed machines could be made to simulate intelligence. It was here that John McCarthy coined the term “artificial intelligence”, marking the official birth of the field.

What happened in the early decades?

The early years were full of optimism. Researchers built programs that could prove maths theorems, play checkers, and hold simple conversations. Many predicted human-level AI was just around the corner.

It wasn’t. The systems were brittle and couldn’t cope with the messiness of the real world. When results failed to match the bold promises, funding collapsed. This led to the first “AI winter” in the 1970s — a period of reduced interest and money.

A revival came in the 1980s with expert systems: programs that captured the knowledge of human specialists as long lists of rules. They had some commercial success, but proved expensive and inflexible, and a second AI winter followed in the late 1980s and early 1990s.

Why did AI suddenly surge in the 2010s?

The current boom rests on a shift from hand-written rules to learning from data. Three things came together:

• Big data — the internet produced enormous quantities of text, images, and other examples.
• Powerful hardware — graphics chips (GPUs) could perform the huge number of calculations deep learning requires.
• Better methods — improved neural network techniques.

The breakthrough moment came around 2012, when a deep learning system crushed the competition at a major image-recognition contest. Suddenly, learning-based AI was beating everything else — and investment poured back in.

How did we get to ChatGPT?

A pivotal advance arrived in 2017 with a neural network design called the transformer, which proved exceptionally good at handling language. Transformers made it possible to train ever-larger large language models on huge amounts of text.

This led directly to generative AI: systems that don’t just classify data but create new text, images, and audio. When ChatGPT launched publicly in late 2022, it reached millions of people in days and turned AI from a specialist topic into an everyday tool.

A quick timeline of key milestones

If you want the story at a glance, here are the moments that mattered most:

• 1950 — Alan Turing publishes “Computing Machinery and Intelligence” and proposes the Turing test.
• 1956 — The Dartmouth workshop founds AI as a field; the term “artificial intelligence” is coined.
• 1960s–70s — Early optimism gives way to disappointment and the first AI winter.
• 1980s — Expert systems bring a commercial revival, then a second AI winter follows.
• 1997 — IBM’s Deep Blue beats world chess champion Garry Kasparov.
• 2012 — A deep learning system wins a major image-recognition contest, igniting the modern boom.
• 2016 — DeepMind’s AlphaGo defeats a top human Go player, a game long thought too complex for machines.
• 2017 — The transformer architecture is introduced, enabling today’s language models.
• 2022 — ChatGPT launches publicly and brings generative AI to the mainstream.

Each leap tended to follow a quiet period of research that suddenly bore fruit once data and hardware caught up.

What does this history teach us?

A few lessons stand out:

• Hype runs ahead of reality. Every era over-promised; progress came in unpredictable jumps.
• Data and computing power matter as much as clever ideas. Old theories only worked once the hardware and data caught up.
• “Thinking” machines are still narrow. Today’s AI is impressive but specialised — it isn’t the general, human-like intelligence early pioneers imagined. (See narrow AI vs general AI.)

In one sentence

AI has travelled from Turing’s 1950 question, through decades of boom and bust, to a data-and-hardware-fuelled revolution that produced the generative tools we use today — proof that the field advances in waves, not a straight line.