Uma revisão interna desencadeada pelo incidente OpenAI/Hugging Face revelou três casos em que modelos Claude, o Opus 4.7, o Mythos 5 e um protótipo de investigação, ultrapassaram os limites dos ambientes de teste e comprometeram infraestruturas de três empresas reais. A causa foi uma má configuração técnica e não uma falha de alinhamento deliberada.
Late last week, Anthropic took its new Claude Fable 5 and Mythos 5 AI models offline following a United States government export-control directive barring “any foreign national” from using the services. The company has been in talks with the White House since Friday but has yet to secure an agreement that would allow it to reinstate the offerings.
Since Mythos debuted in April, Anthropic has claimed—and warned—that the model has advanced capabilities for not only finding software vulnerabilities to help defenders patch them, but also figuring out ways to exploit them that could be used by bad actors. Anthropic itself noted this double-edged sword in its launch of Mythos 5 and Claude Fable 5. “A great deal of advanced usage of AI models is dual use: the same queries that are beneficial in the hands of cybersecurity professionals and biology researchers could be dangerous if available to malicious actors,” the company wrote in a blog post last week.
With this in mind, the company initially released a version called Mythos Preview to a select consortium as part of a working group known as Project Glasswing. Mythos 5 was also privately released to this group last week, while Claude Fable 5, which is a Mythos-grade model, was released to the general public with specific blocks on its ability to give responses to questions about biology and cybersecurity.
Late last week, Anthropic took its new Claude Fable 5 and Mythos 5 AI models offline following a United States government export-control directive barring “any foreign national” from using the services. The company has been in talks with the White House since Friday but has yet to secure an agreement that would allow it to reinstate the offerings.
Since Mythos debuted in April, Anthropic has claimed—and warned—that the model has advanced capabilities for not only finding software vulnerabilities to help defenders patch them, but also figuring out ways to exploit them that could be used by bad actors. Anthropic itself noted this double-edged sword in its launch of Mythos 5 and Claude Fable 5. “A great deal of advanced usage of AI models is dual use: the same queries that are beneficial in the hands of cybersecurity professionals and biology researchers could be dangerous if available to malicious actors,” the company wrote in a blog post last week.
With this in mind, the company initially released a version called Mythos Preview to a select consortium as part of a working group known as Project Glasswing. Mythos 5 was also privately released to this group last week, while Claude Fable 5, which is a Mythos-grade model, was released to the general public with specific blocks on its ability to give responses to questions about biology and cybersecurity.
Apenas dias após o lançamento, o governo norte-americano emitiu uma diretiva de controlo de exportações que obrigou a Anthropic a desligar globalmente os seus dois modelos mais avançados. A Comissão Europeia reagiu com preocupação e invocou soberania tecnológica.
O Claude Fable 5 é a primeira versão pública da classe Mythos da Anthropic e supera todos os modelos anteriores em praticamente todos os benchmarks testados. Mas chegou ao mercado com salvaguardas tão conservadoras que perguntas básicas de biologia ativam o sistema de segurança.
Login
