Anthropic It has reportedly received government approval to restore some access to the Mythos 5 AI model.
This came after an AI startup solved the problem The white house Concerns about the potential risk to national security posed by the Bloomberg News model I mentioned Late Friday (June 26).
“Anthropic has worked with the US government to address risks associated with covered models,” Secretary of Commerce Howard Lutnick He wrote to the company in a letter seen by Bloomberg News. “These efforts have made significant progress,” he wrote, adding that the model could be released to “certain trusted partners.”
According to the report, the cleansing operation calms the conflict that began earlier this month when the government banned entry to Anthropics. Allow foreign citizens And companies are up to Mythos 5 and its sister model Fable 5.
“In just two weeks, we have worked hard to ensure America remains the world leader in artificial intelligence while protecting our security,” Commerce Department spokesman Benno Kass told Bloomberg.
Anthropic issued a statement to Bloomberg confirming that it can begin redeploying Mythos 5 to a small group of cyber defenders and infrastructure providers.
“We are working to get the provider pool approved and restore access to Mythos 5 as quickly as possible. We are pleased to see this progress and continue to work with the government to expand access to Mythos 5 and make Fable 5 available for public use again.”
The report added that it is not clear what measures Anthropic took to allay the government’s concerns could have been done “Escape from prison” Or bypass the guardrails on Mythos and Fable. A source familiar with the matter said the company will have more talks about Fable over the weekend.
Last week, the news outlet spoke with the Executive Vice President of the European Commission, Hina Virkkonen, who said that the regulatory body met with the White House to discuss the matter. Discuss access To Mythos.
In other AI news, PYMNTS wrote last week about how the technology is evolving Corporate expense management More difficult.
This field “has long treated the receipt as proof of the transaction. The employee submits a receipt, a member of the finance team reviews it, and the claim is approved or flagged,” the report said.
This model was fine in an era when creating convincing fake images required photo editing skills or a paid online service, but it no longer works now that free AI image generators can create a receipt with realistic paper texture, fine detail, and a matching timestamp in less than a minute.
“These receipts have gotten so good, we’re telling our customers, “Don’t trust your eyes”” Chris JuneauSenior Vice President and Head of Product Marketing at Saab ConcurPYMNTS said in an October report.
For all of our PYMNTS AI coverage, subscribe to our daily newsletter Amnesty International newsletter.
