AI Timeline

Alan Turing publishes "Computing Machinery and Intelligence," proposing the Turing Test as a measure of machine intelligence. This philosophical foundation asks: can machines think?

The founding event of AI as an academic discipline. McCarthy, Minsky, Shannon, and others gather to explore the possibility of creating thinking machines. The term "Artificial Intelligence" is coined.

Joseph Weizenbaum creates ELIZA, a conversational program that mimics a psychotherapist. Users are surprisingly fooled by its simple pattern-matching responses, revealing human tendency to anthropomorphize.

Unmet expectations and computational limitations lead to reduced funding and interest. AI researchers struggle to deliver on bold promises made in the 1960s, causing government support to dwindle.

Expert systems that capture domain-specific knowledge become commercially viable. Systems like XCON help configure computer systems, generating millions in revenue and reviving AI investment.

Expert systems prove expensive to maintain and brittle. The AI market crashes again as specialized hardware becomes obsolete. Interest and funding dry up once more for nearly a decade.

IBM's Deep Blue defeats world champion Garry Kasparov in chess. This symbolic victory for machine intelligence captures global imagination and signals AI's practical value in complex strategic domains.

IBM's Watson defeats champions Brad Rutter and Ken Jennings in Jeopardy!, a game requiring natural language understanding, knowledge retrieval, and confidence estimation. A watershed moment for AI capabilities.

DeepMind's AlphaGo defeats world champion Lee Sedol in Go. Using deep neural networks and tree search, it masters a game far more complex than chess. A pivotal moment for deep learning and neural networks.

Google releases BERT (Bidirectional Encoder Representations from Transformers). Transformer-based models revolutionize natural language processing, enabling unprecedented language understanding and generation.

OpenAI releases GPT-3, a 175-billion parameter language model that demonstrates remarkable few-shot learning. It can write code, poetry, essays, and more with minimal examples. A leap in general-purpose language AI.

OpenAI releases ChatGPT, an accessible conversational interface to GPT-3.5. It reaches 100 million users in two months—the fastest adoption of any application. AI moves from labs to everyday life.

OpenAI releases GPT-4 with improved reasoning and multimodal capabilities. Anthropic releases Claude, an AI assistant focused on safety and helpfulness. The era of competing large language models begins in earnest.

Advanced models like GPT-4 with Vision and Claude 3 handle images, audio, and documents. Context windows expand to 100K+ tokens. AI assistants become more capable at reasoning through complex, multi-faceted problems.

AI becomes deeply integrated into society—from coding to creative work to scientific research. Questions about alignment, safety, and the relationship between humans and AI systems continue to evolve.