Google's AI Big Sleep discovers twenty new security vulnerabilities in open source
Briefly

Big Sleep, an AI-based vulnerability detector, has uncovered twenty new security vulnerabilities in popular open-source software like FFmpeg and ImageMagick. This effort is part of a collaboration between Google DeepMind and Project Zero, with details of the vulnerabilities currently kept confidential but set to be disclosed to the development community. Earlier in the year, Big Sleep successfully detected a critical zero-day vulnerability in SQLite, preventing its exploitation. Since its introduction in November 2024, it has highlighted the significant role of AI in vulnerability research while Google continues to enhance its AI-driven security tools.
Big Sleep found twenty previously unknown security vulnerabilities in open source software such as FFmpeg and ImageMagick, showcasing AI's impact on vulnerability detection.
The detection of a critical zero-day in SQLite by Big Sleep, before cybercriminals could exploit it, highlights its effectiveness in identifying vulnerabilities.
Read at Techzine Global
[
|
]