Back in December forecasters “estimate(d) a 60% (35% to 75%) probability that the US will have federal or state regulation in force on December 31, 2026 that requires at least some frontier AI developers to publish and follow plans for mitigating the risk that their AIs might cause catastrophic damages.” Why is the probability of the AI Kill Switch Act so low? Do forecasters still think that there is a decent chance of a federal or state regulation being in force on December 31, 2026?
Good question. The AI Kill Switch act is just one proposal, whereas the previous question was about in general. I haven't thought about the federal regulation one much, I'd guess I'd be at about half the original probability. Also note that the executive is taking some faster action here than Congress.
> Some forecasters recommend that readers consider taking the precautionary steps of moving their funds away from European financial institutions, which are going to be more insecure, and into large American institutions, which have access to the latest AI models to faster patch their vulnerabilities with. Our sources tell us that Banco Santander, in particular, is riddled with security gaps, which would make sense in light of its fast expansion.
These recommendations are quite concrete and potentially costly. They imply quite a bold statement (individuals holding funds with American financial institutions are significantly more likely to suffer losses due to cyber then those at European institutions). I'd expect them to be accompanied by relevant forecasts.
What consumer protection laws apply also seems very relevant and not discussed. Claude suggests I am much more protected in the UK than the US. In the UK, the bank is liable and that is enforced by complaint to a regulator, no need to go via courts. The US situation appears less clear cut.
Why moving away from european banks?
Great post! Thanks for the update!
Back in December forecasters “estimate(d) a 60% (35% to 75%) probability that the US will have federal or state regulation in force on December 31, 2026 that requires at least some frontier AI developers to publish and follow plans for mitigating the risk that their AIs might cause catastrophic damages.” Why is the probability of the AI Kill Switch Act so low? Do forecasters still think that there is a decent chance of a federal or state regulation being in force on December 31, 2026?
Good question. The AI Kill Switch act is just one proposal, whereas the previous question was about in general. I haven't thought about the federal regulation one much, I'd guess I'd be at about half the original probability. Also note that the executive is taking some faster action here than Congress.
> Some forecasters recommend that readers consider taking the precautionary steps of moving their funds away from European financial institutions, which are going to be more insecure, and into large American institutions, which have access to the latest AI models to faster patch their vulnerabilities with. Our sources tell us that Banco Santander, in particular, is riddled with security gaps, which would make sense in light of its fast expansion.
These recommendations are quite concrete and potentially costly. They imply quite a bold statement (individuals holding funds with American financial institutions are significantly more likely to suffer losses due to cyber then those at European institutions). I'd expect them to be accompanied by relevant forecasts.
What consumer protection laws apply also seems very relevant and not discussed. Claude suggests I am much more protected in the UK than the US. In the UK, the bank is liable and that is enforced by complaint to a regulator, no need to go via courts. The US situation appears less clear cut.
For bank accounts in the US you are protected by the FDIC.
> American financial institutions are significantly more likely
should be less likely, but yes
> I'd expect them to be accompanied by relevant forecasts.
I agree that it would have been valuable but we prioritized other forecasts. Mildly dislike how this is worded as a demand.
Consumer protection seems important in general but much less important than preserving the assets in the event of chaos.