Presented by
Effective Altruism Finland
We are a Finnish community that asks: what are the world's most pressing problems, and how can we solve them most effectively?
Catastrophic Risks from AI
Registration
About Event
Vili Kohonen, Technical Expert at Tutke, the Finnish Center for Safe AI, will talk about the most severe risks posed by increasingly capable AI systems, from misuse to misalignment. The talk surveys current thinking on existential and catastrophic risks, the empirical evidence behind them, and what is being done to mitigate them.
Mia Hopman, Member of Technical Staff at Apollo Research, will then discuss the concept of scheming in AI agents, introducing how it manifests in today's models, what current detection and mitigation approaches look like, and the open challenges and directions for future work. Mia will be speaking in a personal capacity.
Full series info: acceleratingblind.fi
Presented by
Effective Altruism Finland
We are a Finnish community that asks: what are the world's most pressing problems, and how can we solve them most effectively?