AI-discovered bugs may soon trigger fresh disruptions, cyber experts say
NET•Cybersecurity professionals say AI is accelerating the discovery of software vulnerabilities and could lead to further disruption. Research due to be published by VulnCheck found that digital flaws discovered doubled year-on-year, while Cloudflare CEO Matthew Prince said severe vulnerabilities in other critical software are being patched.
1. More flaws exposed
Seven cybersecurity professionals said advances in AI have uncovered a stream of software vulnerabilities that could trigger disruption. The PHP vulnerabilities behind outages at government websites in Europe and elsewhere this week prompted operators to patch their software, while some took sites offline.
2. Shorter timelines
VulnCheck researcher Patrick Garrity said disclosures involving PHP, Node.js and Plex, along with breaches tied to LiteLLM and Railway, made for a busy week. Experts said AI models can scrutinize codebases continuously and produce exploits in hours, shortening the time defenders have to respond.
3. Experts differ on risk
Experts disagreed on whether more AI-discovered vulnerabilities would lead to more successful hacks. Hunt.io said it saw aggressive attempts to exploit the PHP bug, but its head of research said tracked cybercriminals appeared not to have built a working tool as of 2 p.m. Eastern Time on Thursday. Prince said he had anecdotal evidence of compromises linked to the flaw, but argued that finding and fixing more vulnerabilities would ultimately be positive.
