Featured Image Credit: NurPhoto / Contributor / Getty
AI researchers from some of the world's biggest tech companies have criticised the US government's recent decision to block Anthropic from releasing its new public AI model, Claude Fable 5, warning that the move will be 'dangerous' for the internet.
Claude Fable 5 was one of the most anticipated AI model releases in recent history, as it served as the limited public adaptation of Anthropic's 'Claude Mythos', which was touted as the industry's best cybersecurity tool.
This exclusive model, tested by leading tech companies and even governments across the world, has already proven to be incredibly effective at identifying and dealing with zero day exploits and vulnerabilities, and while Fable offers significant limitations compared to Mythos it still provides game-changing tools to anyone that uses it.
That excitement quickly disappeared just days after the release of Claude Fable 5, however, as Anthropic was informed by the White House that the US government was blocking its release, branding the AI model a 'national security risk'.
Advert
While concerns regarding Claude Fable's potency might be valid – as they have been expressed by Anthropic itself – cybersecurity experts have warned that this could have a significant impact on the internet going forward.
As reported by the Independent, AI researchers and experts from companies such as Adobe, Nvidia, and Zoom have signed an open letter addressed to both Howard Lutnick, US Secretary of Commerce, and Sean Cairncross, National Cyber Director, criticising the Claude Fable 5 ban.
Published on the website 'freefable', the letter asks that the US government "lift the export control directives on Anthropic's Fable and Mythos large language models and commit to an open, scientific and transparent process of handling AI risk assessments in the future."
Anthropic's statement following the ban indicated that it had worked with the US government, amongst other organizations, to red-team Fable's safeguards "for thousands of hours in total," so the block request likely came as a surprise.
What the letter argues is that blocking Fable and Mythos weakens the cybersecurity potential for leading companies in the United States, as it outlined:
"This action has taken the best models away from defenders, created market uncertainty, and risked America's AI leadership without any real risk to justify it."
As mentioned, Anthropic was asked to restrict access to both Claude Fable 5 and Mythos 5 due to unspecified national security risks by the US government, which was born from a request to limit the ability for foreign nations to use the new model.
This directive even included Anthropic employees, and to ensure absolute complicity the AI company decided to simply disable access for all customers to these two specific models.
"Our understanding is that the government believes it has become aware of a method of bypassing, or 'jailbreaking' Fable 5," Anthropic outlined, although it downplayed the possibility of anything harmful actually existing within the tech.
If a universal jailbreak did exist, however, then that could have catastrophic consequences for pretty much every cybersecurity system — which OpenAI CEO Sam Altman has predicted would result in a 'world-shaking cyber attack'.
There would be no better tool to create the perfect attack than the exact same software that is used to cover up any vulnerabilities, but Anthropic seems confident enough in this not being a possibility for the time being despite concern from the government.
There's also the lawsuit currently filed by Anthropic against the Trump administration after the Department of Defense deemed the AI company a 'supply chain risk' for refusing to cooperate with the military's demands, and that might perhaps have been a factor in the government's decision making process.
