Crooks Are Learning to Love AI Hallucinations

Cybercriminals exploit hallucinations in large language models by registering plausible but fake URLs generated by AI, setting up malicious sites to deceive users. Security firms found thousands of such malicious domains and many more unregistered but vulnerable to scammers. This practice, called 'slop squatting,' targets AI-generated recommendations, posing new cybersecurity challenges.