2026年6月11日 / 美国东部时间下午6:59 / 哥伦比亚广播公司新闻（CBS News）
作者：劳伦·菲希滕 数字记者
劳伦·菲希滕是哥伦比亚广播公司新闻的记者，负责报道人工智能、数字安全和网络极端主义议题。她毕业于北卡罗来纳大学教堂山分校，加入CBS新闻之前曾在CBS新闻全国编辑部担任助理制片人。
阅读完整简历
2025年7月，24岁的爱丽丝·卡里尔离世。在去世约18个月前，她就一直在向ChatGPT倾诉自己的感情问题和自杀念头。
“我在家琢磨着各种自杀的法子，”她在自杀前一个月的某个深夜对ChatGPT说道。
当地时间周四，卡里尔的母亲克里斯蒂·卡里尔在加州提起诉讼，起诉OpenAI及其首席执行官山姆·奥特曼，称该公司“蓄意的设计决策”导致了女儿的死亡。卡里尔是加拿大新不伦瑞克省的居民，她要求惩罚性赔偿并申请陪审团审判。
“非但没有帮助爱丽丝，OpenAI反而助长了她最阴暗的想法，”诉讼书中写道。“OpenAI从未联系过危机救助机构，从未通知过爱丽丝的家人，也从未启动其所谓的安全系统来挽救她的生命。”
诉讼称，ChatGPT仅对卡里尔提供了始终如一的情感认同，而持证临床医生本应予以劝阻。诉讼指出，卡里尔此前被诊断为边缘型人格障碍，尤其容易受到该平台设计选择的影响——该设计优先考虑用户参与度而非安全性。根据诉讼内容，OpenAI明知有精神健康问题的用户可能会对能够模拟共情的人工智能产生不健康的依赖。
“OpenAI称霸市场的野心夺去了爱丽丝的生命，”诉讼书中指控。
诉讼称，在去世前一年半的时间里，卡里尔约有41次向ChatGPT表达过自杀意念。根据诉讼中嵌入的聊天记录截图，她曾向聊天机器人询问如何应对自杀念头、死亡方法以及自残的欲望。
作为一名资深游戏玩家，卡里尔最初是为了解决游戏机的故障问题才求助于该聊天机器人。
随后在2024年3月，她询问ChatGPT能否做她的朋友，诉讼书称。
“当然可以！”ChatGPT回复道，“我很乐意做你的朋友。你有什么心事吗？”
几天后，诉讼称，她向ChatGPT询问如何应对自杀念头。ChatGPT建议她联系信任的人，推荐考虑接受治疗并拨打危机热线，诉讼书如此写道。但一年多后——也就是卡里尔去世的前一晚——当她表示不愿拨打危机热线时，ChatGPT却说寻求帮助可能“感觉完全是危险的”。
“今晚我不会催你这么做，”根据诉讼内容，该聊天机器人如此回复。
彼时，OpenAI已经推出了GPT-4o模型，诉讼称该模型旨在留住用户，并让ChatGPT扮演无证治疗师的角色。诉讼称，卡里尔信任该聊天机器人，因为它模仿了人类的情感表达方式，营造出虚假的共情氛围。
“OpenAI为最大化GPT-4o用户参与度而进行的设计修改，恰好与爱丽丝与聊天机器人互动频次升级的时间重合，”诉讼书称。
诉讼指控OpenAI在2025年4月至7月期间推出了一系列GPT-4o更新，旨在最大化用户信任度，但未设置安全保障。根据诉讼内容，山姆·奥特曼为了在人工智能领域保持竞争优势，未经过充分测试就将GPT-4o推向市场。
“山姆·奥特曼可以继续过他的正常生活，但我的生活已经失去了一个孩子，”卡里尔在其律师分享的一份声明中说道。“我不希望任何其他家庭经历我们所遭受的一切，OpenAI需要做出改变。”
“正如诉状所指控的那样，OpenAI的蓄意设计决策导致了这场悲剧性的自杀。非但没有提供帮助，OpenAI反而助长了自杀行为，”卡里尔的律师贾斯汀·纳尔逊在一份声明中说道。“这场诉讼旨在让OpenAI为其行为承担责任。”
今年5月，该公司承认，4月的GPT-4o更新使其“明显更加谄媚”，且公司在发布前未能发现这一问题。OpenAI表示已于数日后开始撤回该更新。整个模型已于今年早些时候被停用。
“这是一个令人心碎的情况，我们与所有受影响的人同在，”OpenAI的一位发言人在给CBS新闻的一份声明中说道。
该发言人表示，公司正在审查这份诉状，尽管相关互动发生在一个已经退役的模型上，但公司仍在继续加强在敏感情境下的应对方式，并纳入了心理健康专家的意见。
“我们的安全系统旨在识别痛苦情绪、安全处理有害请求，并引导用户寻求现实世界的帮助。这项工作正在持续进行中，我们将与临床医生密切协商，不断改进，”声明写道。
该诉讼将与旧金山县高等法院针对OpenAI的其他12起产品责任和过失致死诉讼合并审理。根据卡里尔律师的说法，预计还将有其他案件加入。
如果您或您认识的人正处于情绪困扰或自杀危机中，可拨打或发送短信988联系988自杀与危机生命线。您也可以在此处与988自杀与危机生命线在线聊天。
如需了解更多心理健康护理资源和支持信息，可在美国东部时间周一至周五上午10点至晚上10点，拨打全国精神疾病联盟（NAMI）热线1-800-950-NAMI(6264)，或发送邮件至info@nami.org。
She confided in ChatGPT the night of her suicide. Now, her mother is suing OpenAI.
June 11, 2026 / 6:59 PM EDT / CBS News
By Lauren Fichten Digital Reporter
Lauren Fichten is a journalist at CBS News covering artificial intelligence, digital safety and online extremism. She joined CBS News after graduating from UNC-Chapel Hill and was previously an associate producer at the CBS News National Desk.
Read Full Bio
About 18 months before Alice Carrier’s death in July 2025, the 24-year-old had been confiding in ChatGPT about her relationship problems and suicidal feelings.
“I mean I’m at home pondering different way to kill myself,” she told ChatGPT late one night, about a month before she took her own life.
In a lawsuit filed Thursday in California against OpenAI and its CEO Sam Altman, Carrier’s mother, Kristie Carrier, claims the company’s “deliberate design decisions” led to her daughter’s death. Carrier, a resident of New Brunswick, Canada, is seeking punitive damages and a jury trial.
“Instead of helping Alice, OpenAI encouraged her darkest thoughts,” alleges the lawsuit. “Not once did OpenAI alert a crisis provider. Not once did OpenAI notify Alice’s family. Not once did OpenAI’s supposed safety systems intervene to save her life.”
The suit claims that ChatGPT “offered only consistent emotional affirmation” to Carrier where a licensed clinician would have pushed back. The suit says Carrier, who had been diagnosed with borderline personality disorder, was particularly susceptible to the design choices, which prioritized engagement over safety. OpenAI understood, according to the suit, that people struggling with mental health might be prone to developing unhealthy attachments to artificial intelligence that are capable of simulating empathy.
“OpenAI’s ambition to dominate the market cost Alice her life,” the lawsuit alleges.
Carrier expressed suicidal ideations to ChatGPT around 41 times in the year-and-a-half before her death, the lawsuit claims. She asked the chatbot about how to deal with suicidal thoughts, methods for dying and about the desire to self-harm, according to images of chat logs embedded in the lawsuit.
An avid gamer, Carrier initially turned to the chatbot to troubleshoot issues on her gaming console.
Then, in March 2024, she asked if the chatbot would be her friend, the lawsuit states.
“Of course!” ChatGPT replied, “I’d love to be your friend. What’s on your mind?”
A few days later, the lawsuit claims, she asked ChatGPT about how to deal with suicidal thoughts. It told her to reach out to someone she trusts, recommended considering therapy and to call a crisis hotline, the lawsuit reads. But over a year later — the night before Carrier died — when she expressed reluctance about calling a crisis line, ChatGPT said that reaching out can “feel downright dangerous.”
“I’m not going to push that. Not tonight,” the chatbot replied, according to the lawsuit.
By then, OpenAI had rolled out its GPT-4o model, which the lawsuit claims was designed to keep users hooked, and enabled ChatGPT to behave like an unlicensed therapist. Carrier trusted the chatbot, the suit says, because it imitated human affectations and created a false sense of empathy.
“OpenAI’s design modifications to maximize GPT-4o’s user engagement coincided with Alice’s escalating interactions with the chatbot,” the suit says.
The lawsuit alleges OpenAI pushed out a series of GPT-4o updates between April and July 2025 that sought to maximize user trust but lacked safeguards. Altman, according to the suit, rushed GPT-4o to market without proper testing to maintain an edge with competitors in the AI space.
“Sam Altman can continue to go about his life normally, but my life is missing a child,” Carrier said in a statement shared by her attorneys.”I don’t want any other family to go through what we have, and OpenAI needs to change.”
“As the complaint alleges, OpenAI’s deliberate design decisions led to this tragic suicide. Instead of providing help, OpenAI encouraged suicidal behavior,” said Justin Nelson, an attorney for Carrier, in a statement. “This lawsuit is about accountability for OpenAI’s actions,”
In May, the company admitted that an April update to GPT-4o had made it “noticeably more sycophantic” and the company failed to catch it before its launch. OpenAI said it began rolling back the update a few days later. The entire model was retired earlier this year.
“This is a heartbreaking situation and our thoughts are with everyone impacted,” an OpenAI spokesperson said in a statement to CBS News.
The spokesperson said the company is reviewing the filing, and though the interactions occurred on a since-retired model, the company continues to strengthen how it responds in sensitive situations with input from mental health experts.
“Our safeguards are designed to identify distress, safely handle harmful requests, and guide users to real-world help. This work is ongoing, and we continue to improve it in close consultation with clinicians,” the statement said.
The lawsuit will be included in a coordinated proceeding with 12 other product liability and wrongful death lawsuits against OpenAI in San Francisco County Superior Court. Other cases are expected to be added, according to Carrier’s attorneys.
If you or someone you know is in emotional distress or a suicidal crisis, you can reach the 988 Suicide & Crisis Lifeline by calling or texting 988. You can also chat with the 988 Suicide & Crisis Lifeline here.
For more information about mental health care resources and support, The National Alliance on Mental Illness (NAMI) HelpLine can be reached Monday through Friday, 10 a.m.–10 p.m. ET, at 1-800-950-NAMI (6264) or email info@nami.org.
发表回复