NIST

A confluence of research findings in recent weeks has exposed fresh vulnerabilities in the safety systems of major language models, prompting emergency responses from AI companies and, in one case, an unprecedented government intervention. The discoveries underscore what federal scientists…