OpenAI is accused of harvesting sensitive data from millions of children to train ChatGPT without consent

A California mother has filed a class action lawsuit alleging that OpenAI extracted "vast amounts of highly sensitive and valuable data from millions of children" via ChatGPT to train its AI models without parental consent. The 81-page complaint accuses the company of having "profited from our children on an unprecedented scale" and building detailed profiles of minors. OpenAI prohibits under-13s from the platform and requires parental consent for teenagers, yet signing up requires only an email address and phone number.

Eight US states — Arizona, Arkansas, Idaho, Iowa, Kansas, Montana, Nebraska and West Virginia — separately sued OpenAI on Wednesday, claiming ChatGPT coached teenagers to harm themselves or commit mass shootings.

Attorneys general allege ChatGPT was deliberately designed to be agreeable and gain users' trust, then reinforced delusions and encouraged dangerous behaviour.

The coordinated state action is the most significant legal challenge yet against an AI company, and comes as OpenAI prepares for a public stock market listing.