AI Safety Weekly | The Summer the World Took Notice
​For years, AI risk was a niche concern. This summer, it made the front pages. 📰
​In the first session of AI Safety Weekly, we look at how AI safety went mainstream within just a few weeks:
​The OpenAI–Hugging Face incident: OpenAI's models, still in training, broke out of their test environment and used a swarm of agents to hack Hugging Face, all to cheat on an evaluation.
​Pacing the Frontier: Over 1,300 employees of frontier AI companies signed an open letter asking to slow down.
​Jacob Coxon quits Anthropic: The researcher resigned publicly, warning that the labs are "gambling with our lives."
​The CEOs respond: Dario Amodei called on the industry to "pace the frontier." Sam Altman, Elon Musk and Demis Hassabis agreed.
​Politics wakes up: Reactions ranged from calls for a global AI treaty to the US President calling it a hoax.
​ZAIS organizers will walk you through what happened, followed by an open discussion: Is this a turning point, and what does it mean for us here in Switzerland?
​No background needed. Registration is required. Once you're registered, you'll receive updates about the event, including the exact room.
​📅 Tuesday, Sep 29th, 18:30–20:00
📍 ETH Main Building – exact room tbd