Trump administration allows Anthropic to release Mythos AI to ‘trusted’ U.S. organizations

Source Domain: www.cbc.ca

Anthropic announced that the U.S. government has allowed it to release its powerful AI model, Claude Mythos 5, to over 100 trusted U.S. organizations, reversing a previous order to suspend access over national security concerns.
Mythos 5 now has limited access for companies involved in critical infrastructure security, while efforts continue to expand access and reintroduce the Fable 5 model to the general public.
Concerns remain about the government’s vetting process for which companies gain access to such powerful AI models, raising questions about transparency and rule of law.
There is heightened apprehension that powerful AI could be misused by adversarial military intelligence agencies, prompting the Trump administration’s cautious approach to overseeing AI releases.
An export license is no longer required for Mythos 5 to certain trusted companies and employees who are not U.S. citizens, although strict licensing restrictions are still in effect for entities not on the government’s approved list.
Both Anthropic and its rival, OpenAI, are planning to go public and are navigating government mandates to control the release of advanced AI models to the public domain.
The situation highlights broader concerns about the regulation of powerful AI systems and the potential for these technologies to exacerbate cyber threats if left unchecked.

You may have missed