Anthropic's AI cybersecurity model, Mythos, is now accessible to India and ~150 organizations globally via Project Glasswing. This inclusion, unusual for a non-allied power, highlights India's strategic importance in AI and cybersecurity.
President Trump signed an executive order to bolster US cybersecurity using AI, enhancing federal agency cooperation with AI companies. The order aims to accelerate AI cybersecurity tools and identify advanced AI models, while critics call for stronger, enforceable safeguards.
Anthropic's Project Glasswing, using its unreleased Claude Mythos Preview AI model, has identified over 10,000 software vulnerabilities for partners, significantly increasing bug-finding rates. The initiative aims to prevent AI-driven cyberattacks.
President Trump postponed the signing of an executive order on AI cybersecurity amid concerns it could hinder job growth and due to scheduling conflicts among invited tech CEOs. The order aimed to establish a federal review of advanced AI products before their public release.
Japan's Liberal Democratic Party is proposing enhanced cybersecurity measures in response to Anthropic's powerful new AI model, Claude-Mitos, which poses a potential threat. The party emphasizes collaboration with US tech firms and critical infrastructure sectors.
South Korea's science ministry and other government bodies met with OpenAI to discuss AI cybersecurity. OpenAI shared details on its AI models' security functions and the TAC program, while South Korea requested cooperation on information sharing to combat evolving AI threats.
Finance Minister Nirmala Sitharaman has directed the Indian Banks' Association to help banks counter cybersecurity threats from Anthropic's Claude Mythos AI. Following a Thursday meeting with banking representatives, Sitharaman said banks will discuss investments, technologies and AI use to tackle this emerging threat. SBI chairman Challa Srinivasulu Setty will lead IBA's efforts, with the Ministry of Electronics and IT actively engaging with authorities on preparedness.
Anthropic's new Mythos model has triggered unprecedented global discussion about AI cybersecurity risks, with China explicitly excluded from accessing the breakthrough amid intensifying tech competition between the two powers.
Local banks, power providers and infrastructure firms do not have access to test their systems against a powerful new AI cybersecurity risk, Anthropic's Mythos, which has not been made public.
Anthropic's Mythos Preview AI model helped Mozilla identify 271 security vulnerabilities in Firefox 150, compared to just 22 found by Opus 4.6 in Firefox 148. Firefox CTO Bobby Holley declared that AI tools like Mythos tilt the cybersecurity balance toward defenders, stating 'defenders finally have a chance to win, decisively.' The debate continues over whether such models represent a revolutionary shift in vulnerability discovery or merely incremental AI advancement.
