Maxime Fournes, researcher on AI risk, was a resident at INTP from November 8 to 21, 2023.
He gave a public seminar entitled “Let’s talk about AI extinction risk”, to present the catastrophic risks associated with extremely rapid progress in artificial intelligence, after which Maxime shared the following message:
"[H]ere are some great resources to learn more about AI risk.
* How fast is AI improving? A great interactive showcase of the rate of progress in AI research. It also gives a good intuitive understanding of the dangers of autonomous AI.
* An Overview of Catastrophic AI Risks, a reference paper for the different ways AI can cause a catastrophe, very accessible and covering many different scenarios.
And again, I cannot emphasize this enough, if you are convinced that this risk is real, please consider:
* Being vocal about it around you. The more awareness spreads, the better are the chances that we prevent a catastrophe.
* Joining PauseAI, and generally supporting an international pause on the training of systems more advanced than GPT4."