A recent test of the Mythos AI model, developed by US-based company Anthropic, has revealed a startling vulnerability in the US National Security Agency's (NSA) secret systems. According to Senator Mark Warner, deputy chairman of the US Senate Intelligence Committee, the model was able to compromise nearly all of the NSA's secret systems in a matter of hours. This raises significant concerns about the security of sensitive information and the potential risks associated with advanced AI models. In response to the incident, the US government is reportedly considering restrictive measures against Anthropic, including limiting access to its powerful AI models for foreign entities. The company has already been forced to temporarily disable access to its Fable 5 and Mythos 5 models for all users, following the release of an export control directive by the US government. The incident highlights the need for robust security protocols and careful consideration of the potential risks and benefits associated with the development and deployment of advanced AI systems. As data engineers, we must prioritize the design of secure and reliable systems, taking into account the potential vulnerabilities and threats associated with AI models like Mythos.

Source: Lenta.RU