Phantom Squatting: How AI Hallucinated Domains Become Phishing and Malware

Colleagues, I’d like to flag a growing cybersecurity threat: AI is already creating a new attack vector.

Unit 42 calls it phantom squatting. Models invent non-existent domains, and attackers register them first, then use them for phishing or malware.

In a study, 685,339 queries across 913 brands produced 2.1 million links, 13,229 already malicious, and about 250,000 free domains ready for takeover.

What makes this especially dangerous is that AI assistants and developers often trust such links without verification. This has already led to real phishing and malware campaigns.

The takeaway: treat any AI-generated link as a draft, not a source of truth.

Do you check a domain manually before opening it? #cybersecurity #phishing #AI #threatintel