Anthropic Suspends New AI Tools Amid US Security Concerns Anthropic has temporarily halted the use of its latest AI models, Claude Fable 5 and Mythos 5, following concerns raised by US national security authorities. The company announced the suspension on its website, stating that it had been ordered to restrict foreign nationals from accessing these tools. In a statement, Anthropic described the decision as necessary to comply with the directive, emphasizing that the move was prompted by the government’s belief that the models could be bypassed through a process known as "jailbreaking." This technique allows users to circumvent software restrictions, potentially exposing sensitive data or unlocking features not intended for public access. The company clarified that US authorities had not specified particular vulnerabilities but indicated that a method had been demonstrated to exploit the models. Anthropic acknowledged that the identified issues were relatively simple and noted that other publicly available AI systems could also uncover these weaknesses without requiring bypass techniques. Despite this, the company maintained that its models were "too powerful" to release publicly, a claim that some critics have dismissed as marketing hyperbole. Anthropic had previously highlighted various safeguards to prevent cyberattacks, but concerns from finance, technology, and government leaders persisted, particularly after a private release in April for testing and vulnerability assessments. The suspension comes amid an ongoing legal dispute between Anthropic and the Trump administration. The company is challenging a directive from the Pentagon that labeled it a "supply chain risk," a designation typically reserved for entities based in adversarial nations.#us #pentagon #anthropic #claire_fable_5 #mythos_5
