Speaking at the AWS Summit in Washington, USA on June 30, CIA Director John Ratcliffe compared the capabilities of the most advanced AI models with nuclear weapons, thereby defending the stance of President Donald Trump's administration in controlling the release of the most powerful AI technologies.
In exchanges with many of the President's national security and economic security advisors, we discussed the impact of these advanced AI models," Mr. Ratcliffe shared.
The CIA leader added: "It would not be an exaggeration to consider their capabilities as digital nuclear weapons.
Mr. Ratcliffe said that since taking office about 18 months ago, tracking and protecting emerging technologies has always been his top priority.
According to him, US rivals are seeking to "steal and manipulate US technological achievements for their own benefit", while emphasizing that the CIA is promoting the application of technology developed by US businesses.
The CIA Director also said that the agency has reorganized in the direction of strengthening cybersecurity capabilities, building both "swords" and "shields" to protect key infrastructure. He revealed that he had meetings with Elon Musk, SpaceX leaders, and leaders of Amazon, Google and Dell.
At the conference, AWS - the world's largest cloud computing service provider - announced a $1 billion credit program for US intelligence agencies, and introduced a secure cloud service to serve US defense contractors.
On the same day, Anthropic company said it would soon restore global access to the company's two most powerful AI models, Fable 5 and Mythos 5, after the US government lifted restrictions on where to release these models.
In a post on social network X on June 30, Anthropic said: "We have received notification from the US Department of Commerce that export controls on Claude Fable 5 and Mythos 5 have been lifted. We will begin restoring access from tomorrow.
On June 12, Washington unexpectedly asked Anthropic to stop providing the company's 2 most advanced AI models after discovering loopholes in protection measures to prevent abuse of this technology.
By June 26, the US partially relaxed this order by allowing Mythos 5 to be re-supplied to an approved US partner group, while Fable 5 - a version for general users but still restricted - had not been reopened at that time.
In a letter to Anthropic on June 26, US Secretary of Commerce Howard Lutnick confirmed that Anthropic has coordinated with the US Government to handle risks related to AI models under control.
Like Anthropic, opponent OpenAI also complied with Washington's requirements when releasing the GPT-5.6 model with a very limited scope. "I don't think this is the optimal process" - OpenAI CEO Sam Altman wrote on X on June 26 when explaining the launch of GPT-5.6.
