Trump and AI executives reportedly discussed restricting US AI access based on foreign safety policies

VIEWS1.8KREPLIES1

There have been coherent proposals along related lines btw https://cdn.governance.ai/International_Governance_of_Civilian_AI_OMS.pdf

Though unless I'm missing something OAI has not discussed such things before, and the key first step ofc is the US having real standards, which it does not (besides "have a safety policy")

Miles Brundage@Miles_Brundage

This readout doesn't make sense to me -- why would New Zealand's AI safety policies, which have ~no impact, determine their access to American AI? -- but it's good that OAI is talking about international cooperation at least.

Hopefully just a bad summary

3h1.8K20

LIKES3

Miles Brundage@Miles_Brundage

Child safety is mentioned + it could be the case that New Zealand's laws matter for the impact of American AI in New Zealand.

But that has little to do with frontier models, which is, for now, primarily about what happens in the US.

Unclear if these were treated as one topic

Miles Brundage@Miles_Brundage

There have been coherent proposals along related lines btw https://cdn.governance.ai/International_Governance_of_Civilian_AI_OMS.pdf

Though unless I'm missing something OAI has not discussed such things before, and the key first step ofc is the US having real standards, which it does not (besides "have a safety policy")

3h54630

Cristian Garcia@cgarciae88

@Miles_Brundage I've seen PM meetings like this

3h222

tsunami_crypto@ls_brd

@Miles_Brundage feels like policy theater where the real decision is whoever pays more

hope the full memo clarifies more than this abstract implies

4h5

Hunter Gon@gonlenidefi

@Miles_Brundage Probably means "u must follow our playbook" repackaged as partnership.

Hate when the policy convo skips the real part.

3h