It was John McCarthy. He coined the term in 1955 in his proposal for the 1956 Dartmouth workshop, often seen as the birth of AI as a field.

Turing called it "the imitation game". In his 1950 paper, a judge tries to tell a machine from a human through text conversation alone.

It was Garry Kasparov. Deep Blue beat him in a six-game match in 1997, the first time a computer beat a reigning world champion under standard match conditions.

It stands for Generative Pre-trained Transformer: it generates text, is pre-trained on huge amounts of data, and uses the transformer architecture.

The game was Go. It has more possible board positions than atoms in the observable universe, so many experts thought this win was still a decade away.

It's supervised learning. Each training example comes with the right answer, like photos tagged "cat" or "dog", so the model learns to predict labels.

It introduced the transformer. It uses "attention" to weigh how words relate to each other, and it powers most of today's large language models.

It's called a hallucination. Language models predict likely words rather than check facts, so they can produce fluent but made-up answers.

ELIZA imitated a psychotherapist. Joseph Weizenbaum's program rephrased users' words as questions, and some people got emotionally attached to it.

They were designed for rendering graphics, especially for video games. Their ability to do many calculations at once turned out to be perfect for neural networks.