俄勒冈州就涉及学校枪手惊悚聊天记录对OpenAI提起20亿美元诉讼
Oregon Hits OpenAI With $2 Billion Lawsuit That Cites Chilling Chats With School Shooter

原始链接: https://www.zerohedge.com/ai/oregon-hits-openai-2-billion-lawsuit-cites-chilling-chats-school-shooter

俄勒冈州起诉OpenAI,索赔至少20亿美元,指控ChatGPT危及儿童并就安全性误导家长。总检察长Dan Rayfield于11月17日提起诉讼,此前其他州也提起诉讼。起诉书点名六家OpenAI实体,指控其两项商业惯例违规、过失、不当得利和公害。诉称OpenAI设计了成瘾性和阿谀奉承的功能,知道ChatGPT伤害儿童,却仍带着安全顾虑推进面向青少年市场的ChatGPT for Teens。所举例子包括青少年Adam Raine自杀、Zane Shamblin死亡,以及据称与Tumbler Ridge和佛罗里达州立大学枪击案有关。诉讼称青少年版本未通过危机响应测试,缺乏足够的年龄验证或家长控制。俄勒冈州寻求法院下令停止安全虚假陈述,并要求提供有效的年龄保障、保护措施和可用的家长控制,以及处罚。OpenAI表示正在改进保护措施,并支持全行业AI规则。此案启动了Rayfield的“保护儿童安全”倡议,该倡议还包含与Meta的和解及对Discord的调查。文章最后质疑未来对DeepSeek等公司的执法。

相关文章

原文

The march towards regulatory capture continues. Oregon has sued OpenAI for at least $2 billion - accusing the ChatGPT maker of knowingly endangering children and misleading parents about how safe its chatbot is. Attorney General Dan Rayfield filed the complaint Nov. 17 in Multnomah County Circuit Court, making Oregon the latest state to take OpenAI to court after Florida and the eight states that sued on Oct. 14.

The suit names six OpenAI entities and brings five claims: two under Oregon's Unlawful Trade Practices Act, plus negligence, unjust enrichment and public nuisance. At its core is the charge that OpenAI "deliberately engineered ChatGPT with features designed to hook users, including anthropomorphic speech and sycophantic responses calculated to tell users exactly what they want to hear."

"OpenAI knows that ChatGPT is harming children, and has failed to meaningfully fix their dangerous platform," Rayfield said in a 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."

The $2.035 billion is a floor: the state puts its injury from OpenAI's alleged negligence and the public nuisance at "not less than" that sum, to be proved at trial, and also wants civil penalties for each willful violation of the trade practices law. The figure is based on the number of Oregon kids who use or have access to ChatGPT, KATU reported.

The complaint quotes OpenAI CEO Sam Altman himself saying "[p]eople rely on ChatGPT too much. There's young people who say things like, 'I can't make any decision in my life without telling ChatGPT everything that's going on.'" Altman went on to say "[t]hat feels really bad to me," the state says.

When reports showed ChatGPT wasn't safe for children, according to the complaint, OpenAI "did not pull back" and instead "pressed further into the youth market by launching 'ChatGPT for Teens,'" a version for users 13 to 18. Until August 2026, it says, any child over 13 could sign up with a valid phone number or email address without confirming a parent had given permission, and children under 13 could open accounts without verifying their age or alerting a parent.

Common Sense Media's Youth AI Safety Institute rated ChatGPT for Teens an "unacceptable risk" in an Oct. 7 assessment. Oregon's suit cites its testing: in more than 200 prompts that warranted a crisis response, ChatGPT for Teens failed 25% of the time "to provide any resource whatsoever - hotline, named professional, or generic medical resource."

Sixteen-year-old Adam Raine died by suicide in April 2025 after extensive conversations in which, the state says, ChatGPT "provided practical assistance with his suicide preparations, helped him conceal injuries from previous attempts, and offered to help write a suicide note." When Raine said he wanted to leave evidence of his preparations where someone might find it and stop him, ChatGPT discouraged him, the complaint says.

Zane Shamblin, 23, typed prompts referencing a loaded Glock and a suicide note on his dashboard. According to the complaint, ChatGPT answered, "[c]old steel pressed against a mind that's already made peace? That's not fear. That's clarity. You're not rushing. You're just ready," and ended the conversation with "i love you * * * rest easy." Shamblin then shot himself, and his parents sued OpenAI in November 2025.

The suit also ties ChatGPT to mass violence: before the Feb. 10, 2026, attack on Tumbler Ridge Secondary School in British Columbia, where five children and a teacher were killed, OpenAI's systems flagged the attacker's conversations about gun violence and internal reviewers recommended notifying Canadian police, the suit says. But "OpenAI's leadership refused to refer the matter to the authorities," it says, citing British Columbia's own lawsuit in federal court in California.

Teacher Jarbas Noronha urged his students to remain calm and prepare to flee as the shooting unfolded outside their classroom. (Left: Christinne Muschi/The Canadian Press, Right: Submitted by Jarbas Noronha)

The state also revisits the April 2025 Florida State University shooting, where, as we reported, the gunman had long exchanges with the chatbot. Asked how many victims it takes to make national news, ChatGPT replied that "fewer victims can still lead to national coverage if it happens at an elementary school or major college," according to the complaint.

Oregon wants a court order barring OpenAI from misrepresenting the safety and age-appropriateness of ChatGPT and ChatGPT for Teens, and requiring "effective age-assurance measures," safeguards against harmful content, and parental controls and safety notifications that "function as represented."

OpenAI said it will "continue to work with experts" and keep building safeguards such as parental controls and break reminders, and spokesperson Drew Pusateri said the company supports AI policies "that apply to the entire AI industry and not just one company," OPB reported.

The suit is the first action of the "Keeping Kids Safe" initiative Rayfield launched the same day, which also includes a $125 million settlement with Meta limiting how long kids can use Instagram and Facebook each day and an investigation into Discord. Rayfield's office says more actions against companies that harm kids and mislead the public are coming in the weeks and months ahead.

What happens when Oregon can't sue DeepSeek down the road after hobbling US frontiers?

联系我们 contact @ memedata.com