Dernière minute
Trump and Iran Reach Initial Agreement to End War, Reopen Strait of HormuzUS Government Forces AI Firm Anthropic to Pull Models OfflineXbox Reportedly Closing Ninja Theory StudioUpdated COVID-19 Vaccines Show Significant Cardiovascular Protection, Study FindsMeta's New AI Mode on Facebook Will Use Public Posts for Search ResultsCuban-American Remittance Platform Envioscuba.com Ceases Operations Amid US SanctionsFox Acquires Roku in $22 Billion Deal, Raising Concerns Over Media ConsolidationChinese Rocket Upper Stage Breaks Apart in Orbit, Spreading DebrisNvidia Plans $25 Billion Debt Sale Amid AI BoomAMD Quietly Removes TSME Security Feature from Consumer CPUsTrump and Iran Reach Initial Agreement to End War, Reopen Strait of HormuzUS Government Forces AI Firm Anthropic to Pull Models OfflineXbox Reportedly Closing Ninja Theory StudioUpdated COVID-19 Vaccines Show Significant Cardiovascular Protection, Study FindsMeta's New AI Mode on Facebook Will Use Public Posts for Search ResultsCuban-American Remittance Platform Envioscuba.com Ceases Operations Amid US SanctionsFox Acquires Roku in $22 Billion Deal, Raising Concerns Over Media ConsolidationChinese Rocket Upper Stage Breaks Apart in Orbit, Spreading DebrisNvidia Plans $25 Billion Debt Sale Amid AI BoomAMD Quietly Removes TSME Security Feature from Consumer CPUs
Newsgather

AI Security Institute

Stable2 articles2 sourcesDernière mise à jour: 01.05.2026
Sujets associéscybersecurityanthropicclaude mythoscyberattacksartificial intelligenceopenaigpt-5.5uk government

Derniers articles

UK AI Security Institute Finds OpenAI's GPT-5.5 Can Autonomously Execute Complex Cyberattacks
En développement
Tech·01.05.2026Résumé IA

UK AI Security Institute Finds OpenAI's GPT-5.5 Can Autonomously Execute Complex Cyberattacks

AISI testing shows GPT-5.5 completed a 32-step corporate network attack simulation autonomously and solved a reverse-engineering challenge in 10 minutes that took a human 12 hours. The model achieved 71.4% pass rate on Expert-tier cybersecurity tasks, edging out Anthropic's Claude Mythos. Researchers also identified a universal jailbreak affecting all malicious queries tested.

D
Decrypt
Anthropic's Claude Mythos: The AI That Finds Zero-Day Flaws Raises Global Security Fears
En développement
Tech·23.04.2026Résumé IA

Anthropic's Claude Mythos: The AI That Finds Zero-Day Flaws Raises Global Security Fears

Anthropic unveiled Claude Mythos, an AI model capable of finding and exploiting zero-day vulnerabilities in operating systems and browsers. The company refused public release, comparing it to giving burglars keys to every building. Forty US organizations are partnered in Project Glasswing to patch vulnerabilities before exploitation. Britain tested the model through its AI Security Institute, with ministers warning most businesses remain unprepared. Mozilla's test on Firefox found 10 times more flaws than conventional methods, though none were undetectable by humans. The Pentagon had previously deemed Anthropic a security risk before inviting it back into government cooperation.

G
Guardian Business