OpenAI Sued Over Human Review of User Chats
A federal class action alleges that OpenAI contractors accessed private user conversations without clear disclosure.
Updated on Sept. 24, 2026 in Artificial Intelligence
Live Poll
Should AI companies require explicit user consent before human contractors review private chat conversations?
Two California users filed a class action lawsuit in the U.S. District Court for the Northern District of California alleging that OpenAI contractors reviewed and scored private ChatGPT conversations. The program, known as Project Lily, reportedly allowed human reviewers to summarize and rate model responses.
Why it matters
The complaint centers on whether users were adequately informed that human contractors could access their content for model training and improvement. This case highlights growing legal scrutiny regarding data privacy practices in generative AI model development.
Contractors utilizing specialized dashboards scored model interactions on a scale of 1 to 7. These dashboards reportedly surfaced summaries of previous user activity and location data, while OpenAI retains Temporary Chat logs for safety monitoring for up to 30 days.
The players
OpenAI
An AI research and deployment company known for the GPT series of large language models and the ChatGPT consumer interface.
U.S. District Court for the Northern District of California
The federal court presiding over case number 3:26-cv-10527.
The details
The lawsuit alleges that contractors, recruited via third-party staffing firms, were tasked with reviewing and summarizing user objectives. These reviewers operated through interface dashboards that could display earlier ChatGPT activity along with geographic identifiers. While OpenAI maintains that its consumer-data policy allows for content access to support model improvements unless a user opts out, the litigation challenges the transparency of these data practices.
Timeline
September 16, 2026: The class action complaint was filed in the U.S. District Court for the Northern District of California.
30 days: This is the duration for which OpenAI retains copies of Temporary Chats for safety monitoring.
The Tech Race
This lawsuit follows a pattern set by the 2023 FTC investigation into OpenAI's data privacy practices, signaling increased judicial focus on human-in-the-loop training methods. The outcome could force a broader industry shift in how labs disclose the role of manual data verification in training workflows.
Users concerned about their conversational data can currently opt out of OpenAI's data usage for model improvement through their account settings. This lawsuit may eventually mandate more prominent notifications regarding when and how human contractors interact with private chat logs.
The takeaway
The central issue to monitor is whether courts will define standard human-review processes as a violation of user privacy expectations. Observers should track subsequent filings in case 3:26-cv-10527 to see if the plaintiffs successfully link these review practices to specific contractual or consumer protection breaches.
Further reading
For more on how companies manage the data that powers large language models, see Artificial Intelligence.
Source note: This article includes information reported by TokenPost.
Live Poll
Should AI companies require explicit user consent before human contractors review private chat conversations?
