OpenAI published a proposal Monday asking the US government to lead other countries in setting common standards for the safety and security of frontier AI systems. The plan would have the Center for AI Standards and Innovation work with safety institutes in the UK, France and Singapore on shared ways to evaluate self-improving systems, plus secure channels for passing along emerging threats. Trump turned it down the same day, posting that he won’t “stifle Growth” in a technology he expects to outrun the Industrial Revolution, and that law enforcement can “rein things in if we have to.” Scott Bessent, the Treasury secretary, said last week that labs shouldn’t get a liability exemption, which is how he characterized their ask, and that creators should be liable for what they build.
