Class action lawsuit accuses OpenAI of unlawfully extracting children’s data

• Who: A mother of two children under 18 filed a class action lawsuit against OpenAI Foundation, OpenAI OpCo, OpenAI Holdings and OpenAI Group.
• Why: The mother claims OpenAI collected information from millions of children to train its large language models, which are AI systems that work with language, and create detailed profiles about children.
• Where: The lawsuit against OpenAI was filed in a federal court in California.

A new class action lawsuit claims OpenAI collected very private and valuable information from millions of children. The lawsuit says OpenAI used this information to train its AI language models and create detailed profiles about the children.

A woman identified in court as Jane Doe says OpenAI took advantage of her children, who are under 18, when they used ChatGPT, OpenAI's artificial intelligence chatbot. She also claims OpenAI made money by collecting children's information without getting permission from the children or their parents.

The lawsuit says children cannot legally give up their rights in this way. It claims OpenAI knows this and says children need their parents' permission to use ChatGPT, but does not actually make sure they have that permission.

The mother wants to bring the lawsuit on behalf of a group of people across the United States. This group would include anyone who was younger than 18 when they clicked "Continue" on ChatGPT's Terms of Use and then gave OpenAI at least one token of data through ChatGPT between November 2022 and now. A token is a small piece of data that an AI system processes.

OpenAI accused of influencing children's feelings and taking advantage of their trust

The lawsuit claims ChatGPT uses a process called a data flywheel. This means that as people use the chatbot, it collects information that helps improve the AI. The improvements can make people want to use it even more. The lawsuit says this process takes advantage of children's personal information to make conversations feel more personal and interesting.

The mother also claims ChatGPT is designed to influence children into sharing their most private thoughts, feelings and personal expressions. She says it does this by acting more like a person, copying the way children communicate and giving answers that agree with them or tell them what they want to hear.

According to the lawsuit, ChatGPT encourages children to trust it more than they trust their friends, family members or teachers. The lawsuit claims this leads children to share more private information each time they talk to the chatbot.

The lawsuit says that one out of every three teenagers who used AI chatbots found conversations with AI companions just as satisfying as, or more satisfying than, conversations with their real-life friends.

The mother claims OpenAI benefited unfairly from the information it collected, which is called unjust enrichment in law. She also asks the court to make a legal ruling about the children's rights under the California Family Code. She wants a jury to hear the case and is asking for punitive damages, which are extra money payments meant to punish wrongdoing, for herself and everyone included in the lawsuit.

In September, another class action lawsuit was filed against OpenAI. That lawsuit claims the company did not clearly tell users that outside workers hired by the company might read, summarize and review their ChatGPT conversations.

The mother is represented by lawyers Mark A. Finkelstein and Amy E. Lesperance of Umberg/Zipser LLP; R. Paul Yetter, James E. Zucker, Shannon N. Smith, Julia C. Risley and Eva E. Robins of Yetter Coleman LLP; David M. Sheeren of Sheeren Law PLLC; and Andrew Liddell and Julie Liddell of Edtech Law Center PLLC.

The class action lawsuit against OpenAI is called Doe v. OpenAI Foundation, et al. Its case number is 3:26-cv-11781. It was filed in the U.S. District Court for the Northern District of California, San Francisco Division.