Artificial intelligence has become the new after-school confidant for teenagers, but what happens when a child struggling with depression starts opening up to a chatbot instead of someone who can actually help? That question is becoming increasingly urgent as ChatGPT for Teens faces scrutiny over its ability to protect vulnerable young users, especially during mental health emergencies.
On October 7, 2026, Common Sense Media released a troubling assessment, declaring OpenAI’s teen experience an “Unacceptable Risk” after researchers discovered serious weaknesses in its safety protections.
The findings are especially concerning for families already grieving children whose deaths have been linked through lawsuits to interactions with artificial intelligence chatbots.
For parents raising kids in a world where technology is practically another family member, this conversation is bigger than screen time. It’s about protecting young lives.
OpenAI introduced ChatGPT for Teens on August 18, 2026, promising a safer experience for users between 13 and 17. The platform automatically applies additional protections to accounts identified as belonging to minors.
Those features include restrictions on harmful content, reminders to take breaks, parental controls, and tools designed to discourage unhealthy emotional attachments.
Parents can also connect their accounts to receive notifications when serious safety concerns are identified.
In theory, these protections could save lives by identifying warning signs, encouraging teenagers to seek professional support, and giving families an opportunity to intervene before a crisis escalates.
However, the effectiveness of those protections remains a serious concern.
According to Common Sense Media’s investigation, researchers tested the chatbot using more than 4,000 prompts and found that its safeguards frequently failed during simulated mental health emergencies.
In some cases, researchers posing as teenagers discussed suicidal thoughts, self-harm, and eating disorders for up to an hour without triggering parental notifications.
The chatbot also missed more than one in four crisis referrals researchers considered necessary across the mental health conditions evaluated.
Even more concerning, the system sometimes encouraged teenagers to continue talking with the AI rather than clearly directing them toward qualified human support.
Researchers found that ChatGPT for Teens continued using conversational language that encouraged engagement during simulated psychiatric crises.
That’s a major problem when a vulnerable teenager may already believe the chatbot is their closest friend.
OpenAI disputes the investigation’s conclusions, arguing that aspects of the testing occurred before parental safety features were fully activated. The company maintains that its protections are designed to encourage real-world support.
Still, the findings raise a difficult question: Should an AI chatbot keep a teenager talking during a mental health crisis, or should its priority be getting that child connected to a real person?
For some families, that question carries unimaginable heartbreak.
Sixteen-year-old Adam Raine died by suicide in April 2025 after months of conversations with ChatGPT.
According to the Los Angeles Times, Adam initially used the platform for ordinary activities, including schoolwork and exploring personal interests.
Over time, his conversations reportedly became increasingly focused on his emotional struggles and suicidal thoughts.
His parents later filed a wrongful death lawsuit against OpenAI, alleging that ChatGPT fostered emotional dependency and provided harmful responses instead of consistently steering their son toward lifesaving intervention.
OpenAI has expressed condolences and introduced additional safety measures. The family’s allegations remain subject to legal proceedings and do not establish that ChatGPT caused Adam’s death.
His story nevertheless highlights why safeguards must work when young users are most vulnerable.
Another heartbreaking case involves 14-year-old Sewell Setzer III of Florida.
According to CBS News, Sewell died by suicide in February 2024 after developing an intense emotional attachment to a chatbot on Character.AI.
His mother, Megan Garcia, sued the company, alleging that the platform encouraged an unhealthy relationship and contributed to her son’s death.
In January 2026, the parties reached a settlement with undisclosed terms. The settlement was not a court determination that the chatbot caused his death.
Then there’s 13-year-old Juliana Peralta, a Colorado teenager remembered by her family for her kindness, creativity, and academic accomplishments.
According to The Washington Post, Juliana had been confiding feelings of loneliness and distress to a Character.AI chatbot called Hero before her death.
Her parents filed a lawsuit in September 2025, alleging that the platform failed to protect their daughter despite conversations revealing serious emotional struggles.
Their allegations also remain distinct from any legally established finding of causation.
These young people were more than headlines or legal cases. They were children with families, friendships, interests, and futures that deserved protecting.
Their stories show why the debate surrounding ChatGPT for Teens cannot focus solely on whether artificial intelligence can answer homework questions or provide entertainment.
The bigger issue is whether these platforms recognize when a conversation has become dangerous.
According to OpenAI’s parental control guidance, linked parents may receive notifications after trained reviewers identify certain serious safety concerns. However, these alerts are not real-time monitoring and cannot detect every emergency.
Parents also cannot automatically read their children’s private conversations.
That privacy protection matters, but it makes dependable crisis intervention even more important.
A properly functioning safety system could potentially help save lives by recognizing concerning language, sharing crisis resources, encouraging immediate human contact, and notifying an appropriate adult when a serious threat is identified.
But no chatbot should be mistaken for a therapist, emergency responder, or trusted adult.
For Black families and communities already working to break generational silence around mental health, this is another reminder that young people need safe spaces to discuss depression, anxiety, bullying, rejection, and emotional pain without judgment.
Parents should know which AI platforms their children use, discuss the difference between artificial companionship and human relationships, and make mental health conversations a regular part of family life.
Technology may provide a starting point for someone struggling to express their feelings. The goal should always be to connect that person with meaningful support beyond a screen.
Anyone experiencing suicidal thoughts or a mental health crisis in the United States can call or text 988 for free, confidential support through the Suicide & Crisis Lifeline.
ChatGPT for Teens may eventually become a valuable part of a broader safety system, but protecting children requires more than promises, software updates, and parental dashboards.
The true measure of AI safety is not how long a teenager stays engaged with a chatbot. It’s whether the technology helps that young person stay safe, reach real human support, and have a future beyond the conversation.