Fake Anthropic AI Preview Invites Target Google and Microsoft Accounts

A phishing campaign is using fake Anthropic AI preview invitations to steal Google and Microsoft account credentials. The attack chains together legitimate services and layered CAPTCHA challenges to bypass automated security checks. Interest in AI brands is being exploited as a social-engineering vector.

The fake invitations direct targets through verification steps that appear trustworthy because they rely on real infrastructure and familiar challenge screens. This approach allows the campaign to evade simplistic automated detection while harvesting login details from users who believe they are signing up for early access.

Security teams should note that the credibility of AI brand names is increasingly being used to lower the guard of potential victims. The tactic does not rely on technical exploitation of the AI platform itself but on the willingness of targets to engage with what looks like a legitimate preview offer.