INTERPOL's global security chief told CNBC that AI is accelerating cybercrime by automating the social layer of attacks. Better translation, more convincing fake identities, and agentic AI systems that execute actions on a user's behalf are making scams, fraud, and impersonation faster to produce and harder to detect.
Most companies still treat cyber risk as a network problem. It's increasingly a trust problem. When AI makes phishing emails, fake job applicants, vendor impersonation, and customer-service cons more believable at scale, the weak point shifts from firewalls to the humans and software with real permissions. One polished message and one bad click from an admin can now cost millions.
Security teams will focus on the unglamorous stuff that actually saves money: tighter access controls, stricter approval steps for payments and account changes, and sharper mapping of which systems matter most. INTERPOL's point is that defenses must match the asset and the threat before agentic tools get wider authority inside companies.
The frame that matters: AI didn't create cybercrime. It industrialized the social layer of it. The winners will be security teams that stop treating every inbox, chatbot, and workflow agent as trustworthy by default. The next expensive breach may start with a polished message and one bad click from a user with admin rights.
Filed to the Technology desk · 5 days ago