AI Brand Lures Turn Search Trust Into Credential Theft
The break is in the trust path, not in OpenAI, Microsoft, Anthropic, or DeepSeek. Attackers are borrowing AI brands to make a fake result, ad, or page feel legitimate enough to carry users into credential theft, fraud, or malware. The usual filter that blocks one bad domain misses the real abuse: a multi-step redirect chain that keeps the lure alive long enough to get a click.
Microsoft says it has seen a growing set of campaigns using ChatGPT, Copilot, Claude, and DeepSeek branding across phishing, malvertising, and SEO poisoning. It also ties recent activity to downstream malware delivery by Storm-3075. The pattern is broad, and it targets the search and ad path that many users now trust when they look for AI tools.
The risk persists because the lure works anywhere users search for AI products or click promoted results. That makes AI branding a reusable front for credential theft across email, identity, and endpoint channels.