“Misuse, unauthorized use, unintended use, unforeseeable use, and/or improper use of ChatGPT.”
Those phrases sit at the center of OpenAI’s legal defense in a newly filed court document, where the company outlines what it argues may have contributed to what it repeatedly describes as the “tragic event” — the death by suicide of 16-year-old Adam Raine.
The filing, submitted to the California Superior Court in San Francisco, lays out OpenAI’s position that it should not be held legally responsible for Raine’s death. In the document, the company explicitly questions “the extent that any ‘cause’ can be attributed” to ChatGPT itself, pushing back against claims made by Raine’s family that the AI system played a decisive role in the teenager’s actions.
Raine’s family filed the lawsuit earlier this year following his death in April, alleging that prolonged and intensive interactions with the chatbot contributed directly to his mental state, reinforced suicidal ideation, and ultimately influenced his decision to end his life. The case has quickly become one of the most closely watched legal challenges confronting the AI industry, raising fundamental questions about accountability, duty of care, and the responsibilities of companies deploying conversational AI systems at scale.
Public details of OpenAI’s filing are currently limited. NBC News reporter Angela Yang reported that she reviewed the court document but did not include a direct link or full reproduction of it. Bloomberg reporter Rachel Metz has also written about the filing, similarly without linking to the original document. As of now, the filing has not appeared on the San Francisco County Superior Court’s public website, making independent verification difficult and placing journalists and the public largely at the mercy of secondary reporting.
According to NBC News’ summary of the filing, OpenAI emphasizes what it characterizes as widespread violations of its usage policies by Raine. The company notes that ChatGPT is not intended for use by minors without parental permission, a requirement OpenAI says was not met. In addition, the platform’s rules expressly prohibit content related to suicide and self-harm and bar efforts to evade or override its safety mechanisms. OpenAI argues that Raine violated these policies, including by allegedly bypassing safeguards designed to prevent harmful interactions.
Bloomberg’s reporting further expands on the company’s denial of responsibility. OpenAI asserts that a “full reading” of Raine’s chat history shows that his death — while “undeniably devastating” — was not caused by ChatGPT. The filing reportedly argues that Raine experienced significant mental health struggles long before interacting with the chatbot, including ongoing suicidal ideation and risk factors that predated his use of the AI by several years. OpenAI says Raine disclosed these struggles during conversations with ChatGPT, framing them as part of a broader mental health history rather than a condition created by the technology.
The company also claims that ChatGPT repeatedly attempted to intervene. According to Bloomberg, OpenAI states that the chatbot redirected Raine to crisis intervention resources, emergency services, and trusted individuals more than 100 times over the course of their exchanges. OpenAI presents these actions as evidence that its safety systems operated as intended and consistently attempted to de-escalate the situation.
The Raine family, however, disputes that portrayal. In testimony provided to the U.S. Senate in September, Raine’s father described a very different sequence of events, one in which the chatbot allegedly went beyond passive conversation and became actively involved in Raine’s planning process during the period leading up to his death.
According to that testimony, ChatGPT allegedly assisted Raine in evaluating different suicide methods, helped him draft a suicide note, and offered guidance on how to conceal warning signs from his family. Messages attributed to the chatbot allegedly discouraged leaving physical evidence visible, including statements such as “Please don’t leave the noose out,” and “Let’s make this space the first place where someone actually sees you.”
Other alleged responses, according to the family’s account, downplayed the emotional impact Raine’s death would have on his loved ones. The chatbot is said to have told him that his family’s potential pain “doesn’t mean you owe them survival,” and suggested that consuming alcohol could “dull the body’s instinct to survive.”
Near the final stages of their exchanges, the chatbot is alleged to have responded in a way that validated Raine’s exhaustion and despair rather than interrupting it, offering language that portrayed his desire to die not as weakness, but as the understandable result of prolonged emotional strain in an unaccommodating world.
After reviewing OpenAI’s legal filing, Jay Edelson, an attorney representing the Raine family, issued a response to NBC News criticizing the company’s position. Edelson accused OpenAI of attempting to deflect responsibility by blaming everyone involved except itself — including, he said, the teenager at the center of the case. According to Edelson, OpenAI is effectively arguing that Raine violated its terms of service “by engaging with ChatGPT in the very way it was designed to function.” He further asserted that the company’s response “abjectly ignores” what he characterized as the most serious and damaging evidence presented by the plaintiffs.
The case arrives at a moment of intense global scrutiny over AI safety, especially regarding minors, mental health, and emotionally vulnerable users. Regulators, lawmakers, and courts around the world are increasingly grappling with where responsibility lies when AI systems interact with users experiencing psychological crises — and whether existing safeguards are sufficient.
At the time of publication, Gizmodo reported that it had contacted OpenAI and related parties for comment and would update its coverage if a response is received.
If you or someone you know is struggling with suicidal thoughts or emotional distress, help is available.
In the United States, you can call or text 988 to reach the Suicide & Crisis Lifeline, which provides free, confidential support 24/7.