Microsoft has placed strict limits on its employees' use of Anthropic’s newly released AI model, Claude Fable 5, citing severe data retention and privacy concerns.
The restriction follows Anthropic’s Tuesday launch of Claude Fable 5—a highly capable public version of its cutting-edge "Mythos" AI class. While the model boasts state-of-the-art software engineering and data analytics performance, it comes tied to a strict new compliance policy: Anthropic requires all prompts and generated outputs to be retained for 30 days across every platform to monitor for trust and safety violations. Furthermore, if the system's classifiers flag a prompt as high-risk, Anthropic can store that corporate data for up to two years, News.Az reports, citing Reuters.
According to reports from The Verge, Microsoft’s legal teams are currently scrutinizing these requirements to evaluate the risks posed to proprietary customer data and confidential corporate info.
Because the model forces data outside of standard enterprise security boundaries, it remains unclear if the tech giant will lift the restrictions or clear the powerful model for internal staff use anytime soon.
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。