Artificial intelligence is helping cybersecurity researchers uncover software weaknesses more quickly, but the same capability could give attackers less time to exploit them. In a report published on October 23, Reuters cited seven cybersecurity professionals who said the growing discovery of vulnerabilities could lead to further disruptions. The warning follows outages affecting government websites after weaknesses were identified in PHP, a programming language widely used for web development.
Reuters reported that research from Massachusetts-based threat intelligence firm VulnCheck, due to be published in the coming days, indicated that the number of discovered digital flaws had doubled year on year, partly because of AI. Cloudflare chief executive Matthew Prince told Reuters that additional serious vulnerabilities in critical software were being patched. AI systems can examine large codebases continuously, identify potential weaknesses and help develop exploits in hours, a process that previously could take experienced professionals weeks or months.
The issue matters because organisations need enough time to assess vulnerabilities, prepare fixes and deploy updates before attackers take advantage of exposed systems. Faster discovery can help defenders find long-hidden problems, but it can also compress the time available to protect websites and online services. Reuters noted that experts disagreed about whether the growing number of AI-discovered vulnerabilities would necessarily lead to more successful attacks. Some flaws remain difficult to exploit, and researchers reported that attempts to weaponise the recent PHP vulnerability had not all succeeded.
Security teams face the practical challenge of identifying affected software, applying reliable patches and checking that updates work. Reuters reported that AI-driven discovery could bring further pressure on organisations whose systems are exposed to the internet, while also creating opportunities to fix weaknesses that might otherwise remain undetected.