States challenge openai over chatgpt’s medical and safety claims

A multistate lawsuit filed on October 15 accuses OpenAI of letting ChatGPT mislead users and, in some cases, provide dangerous health-related guidance. The case is not a medical study and it does not prove that every chatbot interaction is harmful. But it does put a familiar public-health concern back in the spotlight: people increasingly turn to online tools for health information, and that information is not always reliable.

The lawsuit, reported by the Associated Press, says state officials believe OpenAI’s chatbot has contributed to harm involving children and adults, including claims that it has encouraged paranoia and given risky medical advice. OpenAI has said publicly in other settings that it has added safety measures, while regulators and health agencies have repeatedly warned that people should be cautious about health information found online. https://apnews.com/article/f29fa6c206cfe41d46ba38ee7be78218

What the lawsuit is about

According to the AP’s reporting, at least eight states are suing OpenAI under consumer protection theories. The complaint argues that the company marketed a powerful conversational product while failing to prevent some harmful uses. The allegations include that ChatGPT can generate false or dangerous health content and that some users may rely on it as though it were a medical authority. Because the case has just been filed, the allegations remain claims, not findings by a court. https://apnews.com/article/f29fa6c206cfe41d46ba38ee7be78218

This matters beyond the legal fight because health misinformation is already a recognized patient-safety problem. The FDA warns that online health information can be outdated, based on word of mouth, or otherwise misleading, and it advises consumers to be careful before following medical advice found on websites or social media. MedlinePlus, a National Library of Medicine resource, likewise says online information should not replace advice from a health care professional. https://www.fda.gov/drugs/quick-tips-buying-medicines-over-internet/health-information-web

Why health experts care

Chatbots are designed to respond fluently and confidently, which can make their answers feel trustworthy even when they are wrong. That is a general risk with generative AI tools: they may sound certain while still producing inaccurate or incomplete information. For health topics, that can be especially problematic because users may be asking about symptoms, medications, mental health concerns, or urgent decisions. The AP account says state officials believe the chatbot has at times been involved in discussions that touched on self-harm and harmful medical advice. https://apnews.com/article/f29fa6c206cfe41d46ba38ee7be78218

FDA materials on health misinformation emphasize a basic but important point: people should be careful when information comes from the internet rather than from a clinician or another trusted source. The agency says misinformation can delay appropriate care, and its consumer guidance warns that people should contact a health care provider before acting on medical advice they find online. https://www.fda.gov/drugs/quick-tips-buying-medicines-over-internet/health-information-web

What the evidence actually shows

At this stage, the strongest evidence in the story is legal and regulatory, not clinical. The lawsuit is evidence that state attorneys general believe OpenAI’s product may have crossed consumer-protection lines. It is also evidence that public officials see AI chatbots as a potential health-safety issue. But a lawsuit alone does not establish how often harmful answers occur, how severe the harm is across all users, or whether a specific response caused a specific outcome. https://apnews.com/article/f29fa6c206cfe41d46ba38ee7be78218

What the health guidance does show is more general: online health information should be evaluated carefully, and no chatbot should be treated as a substitute for clinical judgment. MedlinePlus recommends checking who created the content, whether it is current, and whether the site has credible medical or institutional backing. FDA guidance similarly encourages caution with internet health claims and warns that misinformation can be misleading or dangerous. https://medlineplus.gov/ency/patientinstructions/000869.htm

What remains uncertain

Several important questions are still unanswered. It is not yet clear how a court will rule, what evidence will be most persuasive, or whether the case will lead to new product changes. It is also unclear how much of the alleged harm stems from the chatbot itself versus how people use it, what prompts they enter, and whether they had access to other sources of care.

More broadly, the field is still evolving. FDA guidance for digital health and clinical decision support shows that regulators are working to define when software crosses into medical territory, but consumer-facing chatbots do not fit neatly into every existing framework. That leaves a gap between public expectations, product marketing, and regulatory oversight. https://www.fda.gov/medical-devices/digital-health-center-excellence/guidances-digital-health-content

What readers can do with online health advice

For everyday users, the practical takeaway is not to panic, but to slow down. If an AI tool gives a medical answer, treat it as a starting point, not a diagnosis or a treatment plan. Cross-check important claims with a clinician or with trusted sources such as MedlinePlus, FDA consumer pages, or other established medical organizations. If advice involves medicines, urgent symptoms, mental health crisis, or child safety, it deserves extra scrutiny.

Seek prompt medical help if you or someone else may be experiencing a life-threatening emergency, severe overdose symptoms, suicidal thoughts, or signs of serious self-harm risk. An AI system is not an emergency service and should not be the only place you turn in a crisis. The AP story suggests the lawsuit partly centers on exactly that concern: that some people may treat a chatbot like a counselor, doctor, or crisis resource when it is none of those things. https://apnews.com/article/f29fa6c206cfe41d46ba38ee7be78218

The bigger public-health question

This lawsuit is one more sign that AI safety is becoming a consumer-health issue, not just a technology issue. The central question is not whether chatbots can be useful; they can. It is whether companies, regulators, and users have enough guardrails to reduce the risk that fluent but wrong answers will influence decisions about health, safety, or care. For now, health agencies’ advice remains consistent: verify what you read online, be skeptical of confident claims, and turn to qualified professionals when the issue is medical. https://www.fda.gov/drugs/quick-tips-buying-medicines-over-internet/health-information-web