A newly filed proposed class action lawsuit alleges that OpenAI has been quietly funneling real user conversations from ChatGPT to outside contractors, contradicting the company's assurances about privacy and consent.
The Allegations
According to the complaint, OpenAI operated an internal effort dubbed "Project Lily" that routed genuine user chats to third-party contractors for review. The plaintiffs argue that people who typed private messages, personal questions, and sensitive information into ChatGPT never agreed to have human beings on the outside read those exchanges.
The suit frames this as a breach of trust, contending that users were led to believe their interactions were largely handled by machines rather than seen by human eyes. The heart of the case rests on the claim that OpenAI failed to obtain clear, upfront consent before sharing the material.
The lawsuit claims the people pouring their secrets into a chatbot never knew a stranger might be reading along.
Why It Matters
Human review of AI outputs is a common industry practice used to improve model performance, but the legality and transparency of such programs remain contested. Critics say companies often bury disclosure of these practices in dense terms of service, leaving everyday users unaware that real people may examine their prompts.
The case highlights growing tension between the data appetite of AI developers and mounting public concern over privacy. As chatbots become repositories for deeply personal information, questions about who can access that data—and under what conditions—are increasingly landing in court.
Key issues the litigation raises include:
- Whether users gave informed consent to human review of their chats
- How contractor access to sensitive data was safeguarded
- What obligations AI firms have to disclose data-handling practices
If the class action moves forward, it could set an important precedent for how AI companies must communicate their data practices to millions of users worldwide.