News Summary

Anthropic, a San Francisco-based AI startup, has developed a new artificial intelligence model called Claude Mythos, which has demonstrated remarkable capabilities in identifying software vulnerabilities. Unlike its predecessors, Mythos is being withheld from public release due to fears of misuse by hackers. Instead, Anthropic has partnered with cybersecurity specialists and engineers to utilize Mythos as a defensive tool. The AI model has uncovered thousands of vulnerabilities, some dating back 27 years, which were previously unnoticed. Anthropic has formed an alliance named ‘Glasswing’ with companies like CrowdStrike, Palo Alto Networks, and tech giants such as Amazon, Apple, and Microsoft, to enhance cybersecurity measures. Approximately 40 organizations have joined the initiative, sharing insights and findings from Mythos to protect critical infrastructure. The startup has also engaged in discussions with the US government regarding Mythos, despite facing legal challenges over a White House directive to terminate contracts with Anthropic.