PeterStamp

This is one of several possible ways to get more safety into AI, and some of them will surely be implemented before 2029. Two probabilities to add here for two countries. Generally I´d say, that it´s more likely in the UK, because there the government is more independent from the big tech companies, and closer to more careful EU/Europe. But from recent work for FRI (sic! same letters but different organization) one could see, that the US are also seriously thinking about ways to make AI more controllable (and that would not meet the resistance of Trump´s buddies too much and not weaken the US position compared to China in the AI race too much). I don´t remember this specific one having been on their list of policies they asked us about, but it looks like a good one. One weakness: tax incentives are more relevant, if you actually pay taxes. Start-Ups not yet profitable might not care about them (at least for some ways, in which this could be done).

Good chance, but they might choose other ways. Starting with 15% for UK plus 10% for US.

Will they do something like this? Very likely. Will it fully meet resolution criteria? Less so, but fair chance. Will it happen before 31 July 2026? Unlikely IMO.

"The Trump administration rescinded the Biden-era AI Diffusion Rule in May 2025, just before it was set to take full effect on 15 May 2025 (PBS, CNN). The Commerce Department indicated it would "not enforce the AI Diffusion Rule" and would "issue a replacement rule in the future," with officials calling the Biden rule "overly complex, overly bureaucratic, and would stymie American innovation" (WilmerHale, Mintz). On 21 October 2025, the Department of Commerce established an "American AI Exports Program" in accordance with the Executive Order on Promoting the Export of the American AI Technology Stack, but formal rescission of the AI Diffusion Rule had not been completed (U.S. Department of Commerce, Sanctions News)." (from: "More Info")

time