Linux maintainer Greg Kroah-Hartman on LLM-era security risks
Original: Greg Kroah-Hartman – Security in the LLM Age [video]
Why This Matters
Kernel maintainers setting norms around AI-generated contributions will shape open-source security practices industry-wide.
Linux kernel maintainer Greg Kroah-Hartman delivered a talk titled 'Security in the LLM Age,' addressing how AI-generated code and LLM tools are reshaping threat models and code review challenges in open-source kernel development.
Greg Kroah-Hartman, the long-time Linux kernel stable-branch maintainer, presented a video talk examining security implications brought on by the rise of large language models. The talk, hosted on YouTube, addresses a concern he has voiced publicly before: LLM-generated kernel patch submissions that appear plausible but contain subtle bugs or security flaws — sometimes submitted in high volume by automated pipelines. Kroah-Hartman has previously warned kernel mailing lists about a surge in low-quality, AI-assisted patches that waste maintainer time and introduce review fatigue. In this context, 'security in the LLM age' likely covers both the risk of malicious or careless AI-generated code entering the kernel, and the broader challenge of maintaining trust and signal-to-noise ratio in open-source contribution workflows. No specific new policy announcements are confirmed from the video title alone.