A new SOCRadar report finds over one million infostealer records tied to AI services across more than 80,000 corporate domains, exposing stolen logins and active sessions at 482 large enterprises worldwide. ChatGPT and OpenAI accounts dominate the dataset, reflecting widespread employee use of work emails on personal devices. Stolen session cookies, conversation histories, agents, and embedded API keys give attackers powerful access without needing passwords or bypassing multi-factor authentication. Technology, financial, healthcare, and energy firms all face elevated risk. Researchers urge CISOs to treat AI platforms as core identity systems, enforcing SSO, strict API key governance, session monitoring, and discovery of shadow accounts created outside corporate policy.
This update represents a notable development in the Ai sector. Organizations and founders tracking this space should evaluate potential strategic and technical implications on their operations.