Earlier this week, it was reported that the US government will introduce a new law that will require it to test the latest versions of artificial intelligence (AI) services developed by US AI companies before they are officially offered to the public.
The testing will be conducted by the Center for AI Standards and Innovation (CAISI), which operates under the US Department of Commerce. (US Department of Commerce). According to a Bloomberg report, AI giants such as xAI, Google and Microsoft have agreed to have their AI models tested, while OpenAI and Anthropic have also agreed to be tested after discussions between them and the government.
The agreement means that every major US AI developer will voluntarily participate in the government’s pre-production evaluation. CAISI has tested more than 40 model evaluations so far, including evaluations of AI technologies that have not yet been launched, according to the Commerce Department.
Interestingly, the directive for this pre-production testing was issued by President Donald Trump shortly after Anthropic introduced their latest artificial intelligence model, Mythos, which can reportedly find vulnerabilities in modern software in a short time. Another interesting issue is that Anthropic's artificial intelligence service, Claude, is still blocked and has been labeled a high-risk service by the US government.
