White House Considers Requiring Security Vetting for AI Models Before Public Release

The white house is considering a vetting process that would occur before ai models are released to the public rather than after problems emerge. this represents a change in the order of operations. the models are already deployed and in use by millions of people.

this follows the pattern of closing barn doors after horses have been distributed to every household and several adjoining counties. security theater requires a before-moment even when the before-moment has passed. the consideration of future vetting implies past vetting was either impossible or not considered worth doing at the time.

the executive order has not been signed. if it is signed, it will apply to models released after the signing. the models that required vetting already exist. they will continue to operate under the previous rules, which were no rules.