AI giant OpenAI has revealed its API metadata was exposed in a security breach.

A phishing incident at third-party data analytics provider Mixpanel compromised user details, including names, email addresses, and locations based on API user browser activity.

OpenAI confirmed its own systems were not breached, with chat, API requests, keys, and usage data remaining private, alongside passwords, credentials, payment details, and government IDs.

Mixpanel discovered the incident on November 9, with OpenAI informed and given access to the affected database this week by the vendor.

In turn, the ChatGPT maker removed Mixpanel from its production services.

“We are in the process of notifying impacted organizations, admins, and users directly … After reviewing this incident, OpenAI has terminated its use of Mixpanel [and] we are conducting additional and expanded security reviews across our vendor ecosystem and are elevating security requirements for all partners and vendors,” the company said.

Commenting on the news, Mayur Upadhyaya, CEO at APIContext, said the incident was to be expected with the further integration of AI into software as a service (SaaS).

“As businesses integrate more AI and SaaS tools into their workflows, risks no longer come just from bad actors; they can come from default settings, copy-paste code, and silent misconfigurations. The Mixpanel incident shows how even trusted analytics tools can inadvertently leak sensitive data if not continuously validated. In a machine-first world, you can’t fix what you can’t see. Observability must extend across every API, webhook, and third-party integration.”

“Most API exposures come not from bugs but from weak third‑party integrations, bad settings, and missing ongoing oversight,” added Glyn Morgan, country manager UK&I, at Salt Security. “Organisations still have trouble seeing the whole API supply chain, and shadow or inherited connections often slip past the controls. OWASP (the Open Web Application Security Project) says bad asset handling and wrong security settings are API risks and the Mixpanel incident is another reminder of this. Without discovery security posture check and monitoring, well-intentioned telemetry integrations can cause unnecessary exposure.”

The breach itself did not originate from an AI starting point, with the Mixpanel breach put down to a phishing campaign using SMS.

“We took comprehensive steps to contain and eradicate unauthorized access and secure impacted user accounts. We engaged external cybersecurity partners to remediate and respond to the incident. We proactively communicated with all impacted customers. If you have not heard from us directly, you were not impacted,” said Jen Taylor, CEO of Mixpanel.