Hackers use fake AI and SpaceX invites to steal corporate logins

The lure of exclusive AI access

Security researchers have identified a new phishing campaign that targets the high level of interest employees have in artificial intelligence. Threat actors are creating fraudulent websites for events that do not actually exist, such as an OpenAI Academy livestream or a private SpaceX briefing regarding AI marketing. These lures are designed to appeal to workers who want to stay ahead of the curve in their professional roles.

To entice victims into clicking their links, the scammers offer high-value rewards. These include free ChatGPT Pro subscriptions, access to Grok Heavy, and even the promise of SpaceX shares. By combining the prestige of major tech brands with the allure of free premium tools, the attackers create a compelling reason for users to interact with the malicious content.

As Push Security noted in its analysis, the desire for the latest AI tools makes employees excellent bait for these types of attacks. The campaign specifically targets corporate credentials, aiming to gain a foothold in professional networks. This strategy exploits the current cultural momentum surrounding AI to deliver a highly effective and successful phishing lure.

How the login traps work

The scam functions by directing users to a fake event page where they are asked to log in using their Microsoft or Google accounts. Depending on which provider the victim chooses, the attackers use different technical paths to steal information. For those using Microsoft credentials, the attackers utilize a platform called Kali365 to initiate a device-code phishing attack.

In this specific scenario, the victim receives a device code and is instructed to enter it on a legitimate Microsoft login page. Once the user approves the request, Microsoft issues an authentication token to the attacker's session. The fake event page then redirects the victim to a fraudulent sign-in page operated by VoidProxy, which captures additional credentials and session cookies.

For users who choose to continue with Google, the attackers use a different method involving a reverse proxy. This proxy relays the authentication process while simultaneously capturing the victim's password, multi-factor authentication (MFA) response, and session cookie. This technique allows the hackers to grab an authenticated session without needing to bypass the MFA prompt manually, as they intercept the data in real-time.

Evasion and distribution tactics

To ensure these phishing links reach their targets without being blocked, the attackers use several layers of obfuscation. They utilize Google and Zapier to shorten the links and redirect visitors. These tactics are specifically designed to avoid detection by security scanners and URL reputation systems that organizations typically use to filter out known threats.

While the exact distribution methods are not fully confirmed, researchers assume the links are being spread through phishing emails, social media channels, or Telegram groups. By using well-known services like Zapier for redirection, the attackers can hide the final destination of the malicious URL from automated security tools. This makes it much harder for IT departments to proactively block the scam.

The combination of these evasion techniques with MFA-bypassing payloads means the attack is designed to circumvent the controls most organizations rely on today. Because the attackers can steal session cookies, they can maintain access even if the user has strong password policies. This represents a significant shift toward more sophisticated, automated phishing-as-a-service models.

A growing trend of AI impersonation

This campaign is part of a broader trend of AI-themed phishing that has been escalating over the last few years. Since the first version of ChatGPT was released five years ago, it has been used in countless phishing campaigns. Attackers have moved from simple fake installation pages to more complex lures involving shared conversations and custom GPTs.

Previous reports have shown attackers impersonating brands like Meta, OpenAI, and Anthropic using legitimate Zoom event pages. More recently, researchers observed campaigns using GoTo Webinar to distribute fake invitations. These incidents demonstrate that threat actors are constantly finding new ways to leverage the popularity of AI tools to reach new victims.

As these AI tools penetrate deeper into both private and business lives, they will likely continue to be abused by threat actors. The sophistication of these attacks is increasing as they move toward more seamless ways to steal corporate identities. Security experts warn that the high success rate of these lures makes them a primary concern for modern cybersecurity defenses.