AI finds software bugs twice as fast, changing cybersecurity

Seven cybersecurity experts interviewed by Reuters believe that as AI keeps uncovering hidden software flaws, incidents like this week's government website outages around the world are almost certain to happen again.

Research shows the number of software flaws found in the past year has doubled. This week, government websites in Europe and beyond went down after a security flaw was found in PHP, a widely used coding language. Some sites shut down on purpose to stay safe.

Cloudflare CEO Matthew Prince expects a serious new flaw almost every week for the next two years. AI can scan code around the clock and build attack tools within hours, cutting the time defenders once had to fix things.

Experts still disagree on whether AI will lead to more attacks. Some say AI-found flaws are hard to exploit, and attackers have not yet built a working tool for the PHP flaw. Many believe that in the long run, AI exposing risks does more good than harm.