Google、AIで開発された初のゼロデイエクスプロイトを阻止

google threat intelligence group (gtig) disrupted a zero-day exploit developed with ai that 'prominent cyber crime threat actors' planned for mass exploitation. the python script contained signs of ai generation, including a 'hallucinated' cvss score and code structure mirroring llm training data. it targeted a logic flaw in an unnamed open-source tool to bypass 2fa, though google confirmed gemini was not used. the report highlights a trend of hackers using persona-driven jailbreaking and tools like openclaw to refine ai-generated payloads for increased reliability.