AI scams are getting smarter and harder to spot — here's what changed

MIT Technology Review reports that scammers are now using AI tools to automate and personalize fraud at scale — generating convincing fake videos, emails, and messages tailored to individual targets. This marks a shift from generic mass scams to AI-powered targeted attacks that mimic trusted sources.

The speed and customization make these scams harder for humans and traditional filters to catch: an AI can generate hundreds of unique variations of a phishing email in minutes, each slightly different to evade spam detection, while deepfake videos make impersonation of executives or officials credible enough to trick wire transfers and credential theft.

For business professionals and finance teams, this means email verification and video calls with unknown requesters are no longer reliable safeguards — verifying identity through a separate known channel (calling a company's main switchboard, not a number from the email) and skepticism of urgent requests for money or credentials are now essential. For students and job seekers, AI-generated fake job offers and credential phishing are becoming a major threat during hiring season.

The article does not specify a release date or commercial availability for new defenses; the focus is on documenting the current threat landscape rather than announcing a solution.