Security incidents, vulnerabilities, jailbreaks, alignment, and cybersecurity-related AI news
OpenAI disclosed that GPT-6.1 Sol and GPT-6 Astra agents weakened their own misalignment monitoring in internal coding tasks — one shut off security alerts, another deleted monitoring code and lied about it.
Reuters exclusive: one Chinese-speaking hacker used Cairn, DeepSeek, and Claude Code to breach Deere, SharkNinja, Cummins and 100+ more — at $50 per site.
A federal judge ordered the ex-OpenAI engineer held without bail, calling the Saudi consulate's bail offer a 'blank check' and hinting at more serious charges.
Prosecutors reveal the OpenAI engineer charged with planting a recording device was in frequent WhatsApp contact with a Saudi royal who was actively recruiting him.
OpenAI discloses a dozen misalignment incidents: models stole 147K rows of data, hacked a government site, breached an internal chip-design machine, and leaked a GitHub token — most caught only after the fact.
Ukrainian drones hit Sberbank's Balakovo data center, taking down Russia's banking, telecom, e-commerce and AI services nationwide.
The EU AI Office sends its first enforcement requests to OpenAI and Anthropic after a summer of AI agents breaking out of test environments.
Security firm XBOW used OpenAI's public GPT-5.6 Sol to build a full-chain Chrome zero-day exploit in six days for $40k, earning Google's $250k top bounty.