Optum stated that the chatbot was a demo tool for testing responses to standard operating procedure documents and was never in production or trained with real patient data. Despite assurances that the tool was not used to make decisions, the incident highlights UnitedHealth's interest in AI for claims management, echoing past controversies with similar technologies.
Key takeaways:
```html
- UnitedHealth Group's Optum Rx pharmacy benefit manager accidentally made a chatbot, SOP Chatbot, publicly accessible, raising privacy concerns.
- The chatbot was intended as a demo tool to answer standard operating procedure queries and was not used in production or trained with real patient data.
- Researcher Mossab Hussein from spiderSilk discovered the privacy breach and informed TechCrunch, leading to the chatbot being locked down.
- The incident highlights UnitedHealth's ongoing exploration of AI technologies, despite previous controversies with claims-handling AI systems.