Neural networks first described
Warren McCulloch and Walter Pitts propose the first artifical neuron model — laying the groundwork for the study of neural networks.
Alan Turing envisions human-level machine intelligence
Computer scientist Alan Turing publishes his landmark paper “Computing Machinery and Intelligence,” in which he proposes the creation of a Turing Test to evaluate machine intelligence.
Deep Blue beats Gary Kasparov
World chess champion Gary Kasparov is defeated by IBM’s Deep Blue, and artifical intelligence system capable of evaluating 200 million positions per second.
Autonomous vehicles complete the DARPA Grand Challenge
Five different teams engineer autonomous vehicles which complete a 212-kilometer offroad driving challenge in the Mojave desert, serving as a proof of concept for the viability of self-driving cars.
IBM’s Watson computer defeats human Jeopardy! contestants
A computer leveraging natural language processing techniques was able to defeat human contestants live on air during a taping of trivia game show Jeopardy!
Scientists warn of the dangers of advanced AI
Dozens of the world’s top scientists and leaders, including Stephen Hawking, Steve Wozniak, and Elon Musk, sign a letter warning about the dangers of artificial intelligence.
AlphaGo beats Lee Sedol
AlphaGo captures the world’s attention by beating the former human champion of the board game Go, widely seen as a landmark achievement in human-level strategic thinking.
In 2019, Lee Sedol retired from professional play, citing the strength of AI as his reason.
Transformer architecture invented
The transformer architecture is invented, leading to a proliferation of advanced language models including Google’s BERT and OpenAI’s GPT models.
AlphaStar beats top Starcraft players
Deepmind’s StarCraft-playing AI reaches grandmaster level, making it better than 99.8% of all human players.
AlphaFold is widely recognized for solving the protein-folding problem
For decades, predicting protein structures accurately from amino acid sequences was an unsolved computational challenge, widely known as the protein folding problem. DeepMind’s AlphaFold system made a major breakthrough by leveraging deep learning, dramatically improving the accuracy of protein structure prediction.
OpenAI announces DALL·E
OpenAI announces the release of DALL·E, a new AI system that can make realistic images and artwork from text input. The model demonstrates a sophisticated understanding of language and visual concepts.
NATO announces AI strategy and $1 billion investment
NATO, the military alliance of over thirty countries, announced its 18-point military strategy as well as a $1 billion “future-proofing” fund that will eventually be invested in AI technology with military applications.
OpenAI releases ChatGPT
ChatGPT sets a new precedent for public AI products, growing faster than nearly any internet platform in history by acquiring over 100 million active users in less than two months.
Bing chatbot goes (prematurely) public
Microsoft deploys its new Bing chatbot, internally referred to as Sydney, on the internet. Soon, it’s revealed to have been insufficiently developed and tested for intended behavior — acting erratically and infamously declaring its love for a New York Times journalist and threatening users with blackmail.
Despite this model not having the capacity to carry out those actions, and such behavior eventually being fixed, the premature public release of the model serves as a worrying signal of a lack of corporate oversight.
GPT-4 shows unprecedented capacities
In March 2023, GPT-4 is revealed to the public. It is shown to be capable of passing the bar exam with a 90th percentile score, along with superb performance in other high school- and college-level tests.
It also shows early signs of deception. In the months leading up to release, GPT-4 shows some power-seeking capabilities — including deceiving humans to help pursue its goals.
US Government holds landmark hearings on AI regulation
World governments scramble to make sense of the increasing capabilities of AI, including the United States, where congressional hearings are held to identify crucial considerations for AI regulation.
In the testimony from Sam Altman, OpenAI’s CEO warns that “regulatory intervention by governments will be critical to mitigate the risks of increasingly powerful models.”
Scientists and leaders warn: AI poses risk of human extinction
In an open letter signed by heads of AI labs including OpenAI, Deepmind, and Anthropic, as well as pioneering AI scientists Geoffrey Hinton and Yoshua Benio, the alarm bell is rung. The one-sentence statement reads:
“Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war.”
The future is in our hands.