988 Is Not a Safety Plan

988 Is Not a Safety Plan

Building Safety Plans with Teenagers: A Two-Pronged Approach

I was in a clinic with a teenage patient, building a safety plan. Nothing acute happened that day, that was the point. We were preparing for a night when they might have to face a crisis alone.

When we reached the section on who to contact, I reflexively wrote 988. But upon looking up, I saw hesitation in their face. They knew the number but told me they would never text it because they didn’t know what would happen afterward.

So, we practiced. They sent the first message with me sitting beside them, and I stayed until a counselor responded.

I’m a child and adolescent psychiatrist who works on clinical governance and safety for AI-supported wellbeing tools for young people. I see this problem from two angles: what a young person needs in a crisis and what a technology company should have built before one occurs. I support the crisis-resource requirement in the bill passed by the House. However, I believe it stops too early.

A national survey revealed that nearly three-quarters of teenagers had used AI companions. Among teen users, 33% had chosen an AI companion over a person for important or serious conversations. Character.AI and Google have since faced lawsuits alleging that chatbot interactions contributed to self-harm and deaths.

On June 29, the House passed the Kids Internet and Digital Safety Act, now under review by the Senate Commerce Committee. Its SAFE BOTs provisions mandate that chatbots:

  • Disclose their AI nature at the first interaction and repeat the disclosure if a minor asks.
  • Display resources for contacting a suicide and crisis hotline when a minor discusses suicide or suicidal ideation.

The wording is crucial; it triggers the obligation based on the minor’s prompt, rather than the company recognizing risk indicators. A 15-year-old expressing not wanting to be here anymore might signal danger without explicitly mentioning suicide.

The same provisions would require providers to establish policies to address sexual exploitation, gambling, and substance promotion. However, the bill only mandates a resource for suicide after the young person raises the concern.

Safety Begins Early

The foundation of a safe system is built into the product design. Consider:

  • Data collection: Limiting the amount of information collected.
  • Engagement duration: Encouraging prolonged conversations may not be suitable for minors.
  • Mental health disclosures: Some products may not be designed for such disclosures.

The Youth AI Privacy Act, advanced by the Senate Commerce Committee, addresses these upstream responsibilities. It would limit chatbot data use and restrict features designed to keep young people engaged. It would also fund research and add chatbot use to federal health surveys.

The American Academy of Child and Adolescent Psychiatry recommends testing AI in diverse real-world settings and governing it to ensure automated decisions don’t reduce needed care or override clinical judgment.

A Resource Is Not a Hand-Off

In clinical practice, we don’t let go until we’re certain someone is truly on the other end. A chatbot displaying a hotline number and ending the conversation isn’t enough. It should be a supportive presence, not just a resource.

Leave a Reply

All Rights Reserved. Copyright © AddictionNewsHub.com 2025 AddictionNewsHub.com is not a medical, healthcare or therapeutic services provider and no medical, psychiatric, psychological or physical treatment or advice is being provided by AddictionNewsHub.com. If you are facing a medical emergency or considering suicide or self harm, please call 911 immediately.