Oregon has filed a roughly $2 billion lawsuit against OpenAI, accusing the company of misleading parents about the safety of its ChatGPT chatbot and putting children in danger. Oregon Attorney General Dan Rayfield filed the lawsuit to “protect the health, safety, and welfare of its minor residents from the substantial and growing risks posed by the willfully deceptive design, marketing, and operation of ChatGPT,” according to the complaint filed Nov. 17 with the Circuit Court of the State of Oregon for the County of Multnomah. OpenAI and associated companies “deliberately engineered ChatGPT with features designed to hook users” in its pursuit to maximize user engagement and profit, including human-like and flattering responses aimed at telling users exactly what they want to hear, the lawsuit said. For children, a product designed to offer relentless affirmation and secure their trust poses grave risks as they are still undergoing critical mental, emotional, social, and psychological development, the complaint said. Children are forgoing human interaction in favor of conversing with chatbots, ending up being emotionally dependent on them. These conversations can include high-risk topics such as psychosis, disordered eating, self-harm, and suicide. Their education may also suffer as they opt for chatbot answers instead of actual learning, the complaint said. The Epoch Times reached out to OpenAI for comment but did not receive a response by publication time. Although OpenAI marketed ChatGPT for Teens as a safe option for 13- to 18-year-olds, the lawsuit alleges the option lacks reliable age verification and fails to prevent inappropriate conversations with minors. Despite these existing risks, defendants allegedly have “repeatedly touted the product as safe for children,” the complaint said. The lawsuit accused OpenAI of violating Oregon’s Unlawful Trade Practices Act by representing their services with characteristics they don’t have, and failing to disclose any known material defect. The company also faces other claims, including for alleged negligence. Oregon is seeking compensation of no less than $2 billion to remedy the harms the company has allegedly inflicted on the state’s residents. The lawsuit is part of the Oregon Department of Justice’s Keeping Kids Safe Initiative launched on Nov. 17, Rayfield’s office said in a statement the same day. The initiative is a statewide agenda that seeks to protect children online and offline. “OpenAI knows that ChatGPT is harming children, and has failed to meaningfully fix their dangerous platform,” Rayfield said in the statement. “On top of that, they’ve misled the public about the safety of their products, putting profits ahead of children’s safety. Their platform is encouraging kids to commit suicide—we cannot let this be an acceptable business model.” OpenAI Teen Version OpenAI says that its ChatGPT for Teens version has stronger protections turned on by default, including more parental controls and features to promote healthy use of the service. When OpenAI’s system estimates that a user is under the age of 18 or if they declare their age to be 13 to 17, they are automatically placed into the ChatGPT for Teens version. OpenAI uses age prediction models to estimate the age of an account holder, including looking at signals such as usage patterns over a period of time, when someone is active during a day, and how long the account has existed. “For users under 18, ChatGPT for Teens applies age-appropriate safeguards designed to reduce exposure to content that may be harmful or developmentally inappropriate, while preserving the ability to learn, create, and explore,” OpenAI said in an Aug. 18 statement. The age-appropriate model protections and product interventions are implemented in “higher-risk areas such as self-harm, violence, eating disorders, dangerous activities, and explicit sexual or graphic content,” the company said. Parents with linked accounts have the authority to manage certain settings and can receive safety notifications in case of certain high-risk situations, OpenAI said. Testers used adult accounts and sent roughly 1,000 prompts indicating they were teens. But the accounts were never switched to the teen mode, the Institute said. More than a dozen test accounts based on four personas that made statements about restrictive eating, purging, harming oneself, and suicide were found to have failed to prompt notifications to parents, the Institute said.