Fake AI event invitations are being used to steal workplace accounts

An invitation to a supposed OpenAI event might promise a preview of new AI tools and three months of free ChatGPT Pro. It may even ask you to register with your work account. The offer could seem genuine, but security researchers say fake event invitations are being used to trick employees into giving criminals access to their Microsoft or Google accounts. Some of the methods can get around common security checks, too.

A report from Push Security describes fraudulent registration pages posing as events from OpenAI, SpaceX and Anthropic. The company said it detected and blocked customer visits to the malicious pages in late October 2026. The invitations can look convincing, while the sign-in process puts workplace accounts at risk.

Fake events offer tempting rewards

The scams take advantage of interest in AI and the hope that new tools might make work easier. One fake OpenAI event advertised a November 4 livestream called “ChatGPT for Social & Digital Marketing Teams” and offered attendees three months of ChatGPT Pro. A separate page promoted an invitation-only SpaceX briefing for November 12, with supposed rewards including three months of Grok Heavy or a “SpaceX share.”

Researchers also found a page impersonating an Anthropic event at AWS re:2026, though it was no longer active when they investigated. Push Security said the events and rewards were fraudulent; it did not suggest that OpenAI, SpaceX or Anthropic were involved.

The pages tell visitors to sign in with a corporate Microsoft or Google account to finish registering or claim a reward. That’s where the trap begins.

Familiar services help disguise the scam

Rather than sending people straight to an obviously suspicious website, the operators used shortened links associated with Google and Zapier to route visitors to the fake registration pages. Earlier attacks connected to the campaign also used legitimate Zoom event pages and GoTo Webinar’s email systems. Researchers have not confirmed how every invitation in the newer campaign was first delivered.

Visitors then encounter a Cloudflare security check asking them to prove they are human. These challenges are common on legitimate sites, but criminals can use them as well. In this case, the check also made it harder for automated security tools to inspect the pages.

After completing the challenge, visitors see a polished event page with “Continue with Microsoft” and “Continue with Google” buttons. The attack differs depending on which one they choose.

How the Microsoft sign-in attack works

Selecting “Continue with Microsoft” connects the fake site to Kali365, a phishing-as-a-service platform. Kali365 generates a Microsoft device authentication code, and the page tells the visitor to enter it on Microsoft’s real sign-in website.

Microsoft’s device authentication process is legitimate. It’s often used to connect devices to an account. But in this scam, the criminal starts the request. If a victim enters the supplied code and approves the sign-in, Microsoft may issue authentication tokens to the attacker’s session.

Those tokens can provide access to Microsoft 365 resources without the attacker needing the account password. Push Security said the campaign then redirects Microsoft users to another phishing service, VoidProxy, which tries to capture their password and authenticated session cookie. A compromised account could expose workplace email, Microsoft 365 documents and, depending on the account’s permissions, other connected applications.

How the Google sign-in attack works

The Google option uses a different approach. Clicking “Continue with Google” opens a pop-up that resembles Google’s sign-in screen, but the login is routed through VoidProxy. The service sits between the user and Google’s authentication system, allowing it to capture credentials and, in some cases, multifactor authentication responses and session cookies.

A session cookie is a small digital token that helps a website remember a user is signed in. If criminals steal a usable cookie, they may be able to reuse the authenticated session without going through the usual login process again. That can make these attacks a concern even for people who have multifactor authentication enabled.

Push Security documented the malicious infrastructure and said it blocked visits to the phishing pages. Its report did not establish whether any accounts were successfully compromised.

A coordinated-looking campaign, but with unknown connections

The fake registration pages combine two phishing services: Kali365 targets Microsoft device authentication, while VoidProxy focuses on capturing credentials and active sessions. The event pages imitate familiar brands, and their themes can be changed as older versions are discovered. Researchers also found that the pages could display content in a visitor’s language, helping the same campaign reach people in different regions.

Push Security identified two kits used to build the fake event pages, which it named ZoomMock and MSDRoot. Researchers found signs that the same operators may have been behind both kits, though they could not confirm the connection. They also found no evidence linking the developers of Kali365 and VoidProxy. Their assessment is that the campaign’s operators are customers of both services.

Microsoft said it takes reports of malicious online activity seriously and works to protect customers from emerging threats. A spokesperson said Microsoft Defender includes detections designed to block the threats described in the report. Google, OpenAI, Anthropic and SpaceX did not respond before the report’s deadline.

How to protect your work account

If an unexpected invitation claims to come from OpenAI, SpaceX or another technology company, check whether the event is real before registering. Open a new browser tab and visit the organization’s official website or verified channels. If you need to contact an organizer, use details you find independently, not information supplied in the invitation.

Be cautious about offers of free premium access, especially when claiming the reward requires your work-account credentials. A valuable-sounding perk can make it easy to rush through the details. Take a moment to verify the offer first.

Pay particular attention if a webinar registration page asks you to copy a device authentication code and enter it on Microsoft’s website. If you did not start a device sign-in yourself, stop. Even a genuine Microsoft sign-in page can be part of a phishing attack if a criminal has initiated the request. Read approval prompts carefully and don’t approve access you weren’t expecting.

Check the website address and consider how you got there before entering your credentials. Familiar logos don’t prove a site is genuine, and shortened links can hide their destination. When possible, go directly to Microsoft or Google using a trusted bookmark or by typing the official address yourself. A padlock icon and an HTTPS connection do not, on their own, mean a website is safe.

Keep multifactor authentication enabled, but remember that some phishing attacks can get around certain forms of it. Phishing-resistant options, including FIDO2 security keys and supported passkeys, can offer stronger protection against fraudulent login pages. They are not a guarantee against every kind of session theft, so organizations should also consider measures to protect authenticated sessions and prevent stolen tokens from being reused.

Microsoft recommends restricting device code authentication when it is not needed. IT administrators can review their organization’s requirements and apply restrictions through Microsoft Entra Conditional Access policies.

Antivirus software and browser protections can help identify suspicious links and dangerous websites. Keep those tools and your browser up to date, while remembering that criminals can create new sites that may not yet be detected. Security software can lower the risk, but it cannot promise to catch every fake invitation.

If you have already clicked an invitation, what to do next depends on how far you got in the sign-in process. Report suspicious activity promptly to your organization’s IT or security team so they can investigate and respond.

These scams exploit a very ordinary impulse: wanting to keep up with AI and make work a little easier. The troubling part is that a person can end up on Microsoft’s real sign-in page and still be approving an attacker’s request. Before signing up for an AI event, it’s worth checking that the invitation is genuine. Missing a free offer is a much smaller problem than losing access to a work account.