AI History Timeline

Alan Turing introduces the Turing Test, proposing a machine's ability to exhibit intelligent behavior equivalent to, or indistinguishable from, that of a human.

The Dartmouth Conference, organized by John McCarthy, Marvin Minsky, Nathaniel Rochester, and Claude Shannon, officially marks the birth of AI as a field of study.

John McCarthy develops the Lisp programming language, which becomes the dominant language for AI research.

ELIZA, an early natural language processing program, is created by Joseph Weizenbaum, simulating conversation through pattern matching.

The "AI Winter" begins, as funding for AI research declines due to unmet expectations and difficulties in solving complex problems.

Rise of expert systems in the corporate world, such as MYCIN, designed to diagnose infectious diseases and recommend treatment.

The "Second AI Winter" occurs, as interest and funding for AI declines again due to the limitations of early AI systems and a saturated market for expert systems.

IBM's Deep Blue defeats world chess champion Garry Kasparov, marking a major achievement in AI's ability to handle complex, strategic games.

Geoffrey Hinton and his team at the University of Toronto demonstrate the power of deep learning, reviving interest in neural networks.

IBM Watson wins the quiz show Jeopardy!, showcasing advancements in natural language processing and AI's ability to interpret unstructured data.

AlexNet, a deep convolutional neural network developed by Alex Krizhevsky, wins the ImageNet competition, leading to breakthroughs in image recognition.

Google DeepMind's AlphaGo defeats world Go champion Lee Sedol, an achievement due to the complexity of the game and the use of reinforcement learning.

OpenAI's GPT-3, a language model with 175 billion parameters, demonstrates state-of-the-art performance in various natural language processing tasks.

DeepMind's AlphaFold predicts the 3D structures of almost all known proteins, solving a 50-year-old challenge in biology.

Stable Diffusion 3 adopts rectified flow and the MMDiT architecture, moving the image frontier from U-Nets to transformers.

Black Forest Labs releases FLUX.1, showing a rectified-flow transformer could lead open text-to-image generation.

Physical Intelligence's π₀ puts a flow-matching action expert on a pretrained vision-language model, and robot control becomes a generative problem.

Google's Veo 3 generates synchronized dialogue and sound effects alongside video, in a single pass.

Alibaba's Qwen-Image brings strong text rendering to an open text-to-image model.

OpenAI's Sora 2 ships as a world-model-flavoured video generator with native audio.

The EU AI Act's Article 50 machine-readable marking obligations take effect on 2 August, making provenance a legal requirement for synthetic media.