Cloudflare CEO warns AI will accelerate discovery of critical software vulnerabilities

Cloudflare CEO Matthew Prince has warned that Large Language Models (LLMs) will continue to uncover serious vulnerabilities in critical software for years to come. In an interview with Reuters, Prince noted that both security researchers and cybercriminals are using AI to scan the foundational software of the internet, significantly reducing the time between the discovery of a flaw and the creation of an exploit.

This warning follows recent critical vulnerabilities in PHP, a programming language used by approximately 70% of websites globally. To mitigate risks, several European government websites and the gaming platform GOG were taken offline as a precaution to apply necessary patches. Security groups, including Wiz Research, have already utilized LLM-based tools to identify these flaws.

Despite the increased speed of attacks, Prince expressed optimism regarding the long-term outlook, comparing the ongoing process of securing software to clearing monsters out of a haunted house.