AI will keep uncovering critical software flaws for years, says Cloudflare CEO

Cloudflare CEO Matthew Prince predicts that artificial intelligence tools will continue to discover critical software flaws for years to come. He and his company are already aware of additional serious vulnerabilities in other key software, though he declined to specify which ones. Prince likened these flaws to the PHP bugs that caused numerous websites, including government sites and gaming platforms, to go offline last week.

He expects a period of weekly disruptions as AI enables more rapid discovery and exploitation of vulnerabilities. This shift is attributed to large language models, which can quickly analyze complex code and generate exploits, reducing the time and expertise required. Security researchers and malicious actors alike are leveraging these AI tools to uncover previously unknown weaknesses in the core software that powers the internet.

The Electronic Frontier Foundation reports that over 3% of all software vulnerabilities were discovered in October alone, with nearly a quarter of known security issues reported in the same month. Prince and the EFF warn that this trend will likely worsen before it improves, leading to more frequent breaches, data compromises, and potential service outages.

They advise basic security measures, such as keeping software updated, using strong unique passwords, and employing a password manager. However, Prince remains cautiously optimistic that the cumulative effect of fixing discovered flaws will eventually reduce the overall vulnerability of the internet's core software.