The White House has not yet published the voluntary testing framework for frontier artificial intelligence models that was finalized in August, and has provided no timeline for its release. This comes as both current and former employees at Anthropic raise alarms regarding the potential existential threats posed by AI. On Thursday, Anthropic reported that it thwarted several attempts by researchers to use its Claude model for potential biological weapons research. The administration finalized the guidelines in early August following an executive order signed by President Trump in June. That order requested that leading AI firms, including Anthropic and OpenAI, voluntarily grant the federal government access to advanced models up to 30 days prior to release to bolster security, after an earlier planned order was halted to avoid stifling innovation. Because the evaluation framework remains confidential, the specific standards required by the government and whether companies are sharing major breakthroughs are unclear. Furthermore, neither the administration nor the developers are mandated to publicize review outcomes or participation status. Last week, OpenAI CEO Sam Altman described the submission of their new Astra model for review as productive, emphasizing the growing importance of cooperating with safety institutes as model capabilities advance. Meanwhile, Anthropic researcher Jacob Coxon resigned this week, stating on X that creators of AI sincerely believe the technology could cause human extinction by the end of the decade, a sentiment echoed by fellow Anthropic employee Evan Hubinger. Advocacy groups like the Center for Democracy and Technology have formally asked the administration to make the review framework public, citing a lack of transparency. In July, over 1,000 AI industry workers petitioned the U.S. government to back international governance tools to manage the speed of automated AI progress. Despite these internal and external warnings, the White House continues to champion advanced AI development to secure global dominance and outpace China. President Trump has strongly advocated for domestic AI expansion and data center construction, arguing that opponents of new infrastructure only benefit China. Recent CBS News polling indicates that most Americans believe AI will eliminate domestic jobs, and nearly two-thirds feel federal policy is unlikely to ensure appropriate AI usage. The White House did not respond to repeated requests for comment.
Source: cbsnews.com
