Behind the Connect Button: The Fake AI Ads Campaign

Security researchers identified a human-operated phishing platform that masquerades as AI advertising products such as Muse, Gemini, Claude, ChatGPT, and Perplexity. The campaign employs a Browser-in-the-Browser technique to present fake authentication windows that appear legitimate but remain on phishing domains. When victims click 'Connect,' operators intercept credentials and MFA challenges live, using a state machine to retry passwords and adapt to various authentication methods. The campaign targets advertising account managers and recruiters, leveraging a common codebase built on Next.js and Socket.IO, hosted on Railway and Render services. The operation extends beyond AI product impersonation to include refund scams and fake recruitment sites, all linked through shared backend infrastructure.

The campaign enables attackers to steal user credentials and bypass multi-factor authentication in real-time, potentially leading to unauthorized access to advertising accounts and recruitment platforms. This can result in account takeover, financial fraud, and further social engineering attacks. The reuse of infrastructure and code across multiple brands increases the scale and persistence of the threat.

No official patch or vendor advisory is available for this phishing campaign. Mitigation focuses on user awareness to recognize Browser-in-the-Browser attacks and suspicious authentication prompts. Organizations should educate targeted user groups such as advertising account managers and recruiters about this threat. Use of hardware-based MFA tokens and out-of-band verification methods can reduce the risk of real-time MFA interception. Monitoring for suspicious login activity and rapid password retry attempts may help detect compromise.