Skip to content
Submit incident
AII-1002-55283
Well documentedThree or more independent publishers report the core facts.Supporting sources are pending verification.

Optum AI chatbot publicly accessible without authentication

Optum's internal AI chatbot for handling employee questions about insurance claims was accidentally exposed on the public internet without password protection. A security researcher discovered the exposure, and Optum restricted access after being contacted by TechCrunch; the company stated no patient data was compromised.

13 December 2024 Β· Optum Β· Security and cyberattacks
13 December 2024
AII-1002-55283
Quick facts
Category
Security and cyberattacks
Country
United States
Date
13 December 2024
Drafted by
Raidu Incident Detection

What happened

Optum's internal 'SOP Chatbot,' an AI tool designed to help employees answer questions about insurance claims and company standard operating procedures, was publicly accessible on the internet without requiring authentication. The chatbot was hosted on an internal Optum domain, but its IP address was publicly discoverable and accessible via web browser to anyone who had it, without password protection.

Security researcher Mossab Hussein, chief security officer of cybersecurity firm spiderSilk, discovered and reported the exposure to TechCrunch in December 2024. After TechCrunch contacted Optum for comment, the company restricted access and the chatbot is no longer publicly accessible.

Reported impact

Optum stated that the chatbot did not contain or produce sensitive personal or protected health information, and no real patient data was inserted into it or used in its training. The company characterized the tool as a 'demo' and 'proof of concept' that was 'never put into production' and 'never scaled nor used in any real way.'

Index analysis

The chatbot's training data consisted of internal Optum documents relating to standard operating procedures for handling claims. Employee query logs were visible and accessible, showing staff asking questions such as 'How do I check policy renewal date?' and 'What should be the determination of the claim?'

Governance controls

The tool was made public without authentication controls despite being an internal system hosted on an Optum domain. The duration of public exposure is unknown.

Classification

Category
Security and cyberattacks
Industry
Healthcare, drugs and biotechnology
Country
United States
Harm status
Potential harm
Intent
Unintentional
Timing
Pre-deployment
Harmed
Workers
Failure modes
Data leak
Consequences
None recorded
AI systems
SOP Chatbot

Sources and evidence

Sources not yet published

This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.

How sources are verifiedSuggest a source for this incident