NewsIncident
Push Security: fake OpenAI, SpaceX and Anthropic event invitations steal Microsoft and Google work accounts with device code and relay phishing
Push Security's Luke Jennings reported on November 5, 2026 a campaign of fake registration pages for AI-themed events: OpenAI Academy livestreams, a private SpaceX "AI for Marketing" briefing and an Anthropic session at AWS re:Invent, with rewards such as three months of ChatGPT Pro. The invitations use trusted redirectors, and a Cloudflare challenge keeps scanners out. Each page asks for a work sign-in. The Microsoft button runs device code phishing through the Kali365 platform, which gives the attacker tokens without the password, and then sends the victim to a VoidProxy page that captures the password and session cookie; the Google button opens a VoidProxy relay of Google's real sign-in that captures the password, MFA response and session cookie. Push says the operator moved from Zoom lures to Anthropic, then OpenAI and SpaceX, between October 20 and 30, and believes it is a customer of both services. BleepingComputer, which also covered it on November 5, reports no victim count and says the first delivery step is unclear.
Why it matters
An invitation to an AI event is the bait, and the real target is the session after you sign in. Block or limit device code sign-in where it is not needed, and move to phishing-resistant sign-in such as security keys, which a relay page cannot replay.