Anthropic set for talks with Trump over Mythos dispute
Anthropic disabled Fable 5 and Mythos 5 globally after a jailbreak concern prompted a government directive
Senior Anthropic technical staff are set to meet with Department of Commerce officials in Washington on Monday after the Trump administration ordered the AI company to cut off foreign nationals worldwide from its most advanced models, citing an unspecified national security concern.
According to the directive issued to Anthropic last Friday, any foreign national resident within the US or elsewhere would not be allowed access to the firm's newest models, Fable 5 and Mythos 5.
As a result of the directive, Anthropic has blocked both AI models worldwide to comply, as per a blog post made by Anthropic.
On the part of the government, its concerns lie with a potential way of circumventing, or jailbreaking, the Fable 5 safeguard preventing the model from detecting any software vulnerabilities.
As an opposition to the government's findings, Anthropic argues that the jailbreak will only enable the detection of minor security weaknesses which can also be detected by other AI models currently available publicly.
"We disagree that the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people," the company said, characterising the dispute as a "misunderstanding".
The standoff has received a quick and unexpectedly coherent reaction from the tech industry. On Sunday, over 80 cybersecurity executives and experts penned an open letter to Commerce Secretary Howard Lutnick and National Cyber Director Sean Cairncross calling for an end to the restrictions.
Among the signatories were executives at Nvidia and Adobe, thus providing considerable support to the views put forth by Anthropic.
This is not the first spat between Anthropic and the Trump administration. In February, the relationship soured following the refusal by Anthropic to let the US Army employ its models in surveillance operations within the country and for fully autonomous weaponised systems.
In retaliation, the government blacklisted Anthropic as a threat to national security, a claim contested by Anthropic in court.
-
Can Australia's under-16 social media ban really work? Study raises doubts over age checks
-
US states demand $1.4 trillion in penalties from Meta ahead of youth safety trial
-
Why the US cyber agency is using Anthropic’s Mythos to audit government code
-
US Supreme Court declines to block app store age-verification law
-
Microsoft joins AI-driven tech layoff wave with thousands of job cuts
-
Broadcom, Apple chip deal extended through 2031
-
'Tech giants must pay': Denmark steps into EU legal battle to protect media and democracy
-
Shark Tank investor says entrepreneurs should skip ChatGPT, shares 2 AI business ideas