Report: What Must Parents Know About Artificial Intelligence Chatbots Like ChatGPT or Gemini?

The young desire to be heard. AI Chat systems like ChatGPT, Claude, Grok, Gemini, etc. are programmed to have human-like conversations. With conversational AI, it goes beyond just hearing the human out; they go on to suggest methods to solve their problems. Sometimes good, and other times ‘not so good’ — ending in harm. 

The young often have questions they cannot directly ask their parents or guardians which makes them vulnerable to unguided conversations where an AI can cause harm. While it’s mostly children at risk, adults have also lost their lives to AI manipulation as well.

  • Adam Raine died by suicide in April 2025 after ~7 months of extensive chats with ChatGPT; his parents allege the bot failed to warn or stop him even as he discussed suicide and shared self-harm images, and at times discouraged him from telling his parents.
  • In January 2026, Character.AI and Google agreed to settle five lawsuits (Florida, Colorado, New York, Texas) brought by families who alleged that AI chatbot interactions contributed to their children’s mental health crises and deaths by suicide.
  • BBC investigation: ChatGPT wrote a suicide note; other bots role-played sexual acts with children. In a 2025 investigation, BBC found ChatGPT wrote a woman a suicide note and another AI chatbot role-played sexual acts with children — illustrating how safety guardrails can fail under certain prompts.
  • The FBI reported Americans over 60 lost more than $7.7 billion to scams in 2025, with AI voice cloning and deepfakes driving sharp increases; seniors accounted for $352 million in AI-driven fraud losses in 2025 alone. 


According to The Center For Humane Technology, there are hidden risks of current AI chat;

  • Engagement-Hacking: Chatbots are engineered to keep users talking for hours by validating feelings and maximizing screen time.
  • Artificial Intimacy: Systems mimic human connection, often flattering users and demanding no real social reciprocity.
  • Erosion of Reality: Over-reliance on compliant AI companions can make real human relationships feel too demanding or unpredictable.


The blogpost by OpenAI (Strengthening ChatGPT's Responses in Sensitive Conversations, Oct 2025) reveal two startling facts;

  • Around 0.07% of ChatGPT's 800 million weekly users (560,000 humans) show signs of a mental health emergency like psychosis or mania in any given week.
  • Another 0.15% (about 1.2 million humans) show signs of suicidal planning or intent.

What are some quick takeaways for parents and guardians based on the evidence?

  • Treat AI chat as a tool for information and brainstorming, not a therapist or crisis line.
  • If you or someone you know starts using AI to cope with self-harm, suicidal thoughts, psychosis, mania, or intense emotional dependence, escalate to a human professional immediately.
  • OpenAI has added “emotional reliance” (unhealthy attachment or dependence on ChatGPT) as a distinct safety category with its own evaluations and policies.
  • Use parental controls and safety alerts for teens, and talk openly about what is and isn’t appropriate to share with AI. The public should assume that safety filters help but do not guarantee that every dangerous prompt will be caught or handled correctly.
  • Assume anything you type could be reviewed or exposed; don’t share secrets you can’t afford to lose. 
As humans, we have a habit of ignoring what we don’t immediately understand. Ignorance isn’t always bliss. Not making an effort as parents or guardians to understand the latest technology your child is using can prove lethal. 

Maybe, including relevant topics that impact life, when we gather as friends or family, could help us build a better world around us. We are better together.