Cybersecurity researchers uncovered a phishing platform that mimics advertising portals for major AI chatbot services. The fake sites impersonate campaign management and business account tools for Google Gemini, Anthropic Claude, OpenAI ChatGPT, Perplexity, Meta Muse, and Manus.
The phishing operation uses deceptive advertising portals that claim to offer campaign optimization, spend audits, and business-account connections. Researchers describe it as a "human-operated phishing platform" designed to harvest sensitive information from users who believe they are accessing legitimate ad management tools.
The attack captures two critical authentication elements. Users unknowingly enter their credentials on the fake portals. The platform also collects multi-factor authentication (MFA) codes, which attackers can use immediately to access victim accounts even if the user has enabled two-factor protection.
The breadth of AI services targeted suggests the attackers cast a wide net across the AI sector's user base. Marketing professionals, business account managers, and campaign operators represent the likely targets, as they would naturally seek out ad optimization tools for these platforms.
Researchers attributed the disclosure to The Hacker News. The investigation revealed operational sophistication in the phishing infrastructure, with custom-built fake portals designed to closely mimic authentic interfaces. The use of "human-operated" tactics indicates active management and ongoing refinement of the attack strategy rather than automated credential harvesting.
This phishing campaign demonstrates how threat actors exploit the rapid expansion of AI services and the complexity of managing multiple advertising accounts across different platforms. Users managing ads for AI chatbots face increased risk from credential theft operations that target the intersection of business operations and AI service adoption.
Security measures like MFA typically protect accounts, but this operation's ability to capture and use MFA codes in real time bypasses that protection layer. Organizations using these AI chatbot