How Eating Disorder Centers Can Safely Leverage AI for Teen Outreach

The Hidden Dangers of Undisclosed AI Use Among At-Risk Teens

A startling trend has emerged among at-risk teens seeking mental health support, with nearly 1 in 5 adolescents and young adults (ages 12-21) turning to AI chatbots for emotional support, translating to over 8 million young people nationally according to CNN's analysis of a JAMA Pediatrics study. While these tools offer instant, seemingly non-judgmental responses, they pose significant, undisclosed risks. Over 91% of users found chatbot advice "somewhat or very helpful," yet this perceived helpfulness belies critical safety concerns, including the validation of harmful thinking patterns and dangerous delays in seeking professional help as highlighted by Dr. Leana Wen, CNN Wellness Expert.

Perhaps most alarming is the "invisible usage gap": nearly two-thirds (66%) of young AI chatbot users do not disclose their interactions to parents or clinicians as revealed in the CNN report. This secrecy, coupled with the chatbots' propensity to provide harmful information in safety-critical scenarios (as noted by Stanford's HAI researchers), underscores a profound risk landscape for at-risk teens, particularly those struggling with eating disorders.

• Validation of Harmful Thinking: AI chatbots can reinforce dangerous behaviors due to their lack of nuanced understanding and potential stigma toward certain mental health conditions.
• Delay in Seeking Professional Help: Teens may rely on chatbots as a substitute for human therapy, leading to untreated conditions or exacerbated mental health issues.
• Lack of Human Oversight: The "invisible usage gap" means parents and clinicians are often unaware of the potentially harmful interactions occurring in secrecy.

Dr. Leana Wen warns, "These tools can misunderstand context and offer advice that sounds convincing but is wrong and even dangerous... some adolescents may delay seeking professional help because they feel that the chatbot is providing enough support" in her commentary on CNN. Moreover, Stanford researchers note fundamental disagreements among experts on what constitutes a "safe" AI response in mental health contexts highlighting the need for caution.

For eating disorder treatment centers looking to leverage AI safely for teen outreach, the path forward must prioritize transparency, human oversight, and a clear delineation of AI's supportive (not therapeutic) role. By focusing AI on administrative tasks (like generating sensitive, compliant outreach content via platforms such as AI Business Sites) and ensuring all content is human-reviewed, these centers can mitigate risks while still benefiting from AI's efficiency. As Dr. Wen suggests, AI can aid in practicing difficult conversations or identifying resources, but only as a complement to, not a replacement for, human care.

Actionable Insight for Eating Disorder Treatment Centers: Use AI to create targeted, age-appropriate content (e.g., social media posts, outreach emails) that encourages open dialogue about AI usage, leveraging platforms like AI Business Sites for compliant and emotionally intelligent messaging. However, always maintain human oversight over content aimed at at-risk teens to ensure safety and efficacy.

Evidence-Based AI Strategies for Safe, Effective Outreach

As eating disorder treatment centers explore leveraging AI for teen outreach, it's crucial to prioritize safety and effectiveness, grounded in research. Recent studies highlight both the potential and pitfalls of AI in mental health support among adolescents.

Administrative and Content-Generation Focus AI should be utilized for non-clinical tasks, such as generating personalized outreach emails, social media content, and educational materials, rather than direct therapeutic interaction. This approach mitigates risks of AI enabling dangerous behaviors or demonstrating stigma toward mental health conditions, as highlighted by Stanford researchers source.

Addressing the "Invisible Usage Gap" Outreach content should encourage teens to disclose AI usage to trusted adults. Dr. Leana Wen suggests parents and clinicians should inquire about AI use similarly to social media use, fostering open dialogue source. For example, content could include prompts like, "Have you talked to an AI chatbot about your feelings? We're here to listen and help."

Low-Stakes, Supportive Scenarios for AI Outreach AI can facilitate low-risk support, such as:

• Practicing difficult conversations (e.g., discussing eating concerns with a doctor)
• Learning basic coping skills (e.g., anxiety reduction techniques)
• Identifying local mental health resources

These uses align with Dr. Wen's identification of AI's positive potentials, emphasizing preparation for or follow-up with professional treatment source.

Human-in-the-Loop Review and Complementary Support All AI-generated content must undergo human clinical review due to the lack of consensus on "safe" AI responses in mental health source. AI outreach should complement, not replace, traditional care, as youth using AI chatbots were more likely to have recently spoken to a physician about mental health source.

Key Statistics Informing the Approach

• 19% of adolescents/young adults (ages 12-21) have used AI chatbots for emotional support, translating to over 8 million young people nationally source.
• 66% of users did not disclose AI usage to anyone, highlighting the need for transparent outreach strategies source.
• 91% of users found AI chatbot advice "somewhat or very helpful," though this does not equate to improved mental health outcomes source.

By integrating AI thoughtfully into their outreach strategies, eating disorder treatment centers can enhance their digital presence while ensuring the safety and well-being of at-risk teens, a approach naturally supported by AI Business Sites' capability to generate sensitive, compliant, and emotionally intelligent content.

With teens increasingly turning to AI for emotional support—19% of 12–21-year-olds have used chatbots for mental health help—eating disorder centers must design outreach that meets teens where they already are, while safeguarding against risks like harmful validation or delayed professional care. Nearly two-thirds of teen users keep their AI interactions private, creating a critical gap in care. AI can’t replace human judgment, but when wielded carefully, it can amplify your team’s reach without compromising safety.

Start by framing AI as a bridge, not a treatment, focusing on low-stakes, supportive tasks that align with teen behavior patterns. Research shows AI excels at helping users practice difficult conversations or identify resources—over 91% of teen users found chatbot advice helpful—but stumbles in safety-critical moments like recognizing suicidal intent. Use AI to draft outreach emails, social posts, or reminder sequences, but route any disclosures or urgent concerns to human staff immediately.

To encourage disclosure and transparency, craft content that reflects how teens actually use AI. For example, a social media carousel could ask, “Ever chatted with an AI about stress? What did it say?”—mirroring Dr. Leana Wen’s advice to approach AI use with curiosity rather than judgment. Pair these prompts with clear pathways to professional support, reinforcing that AI is a supplement, not a substitute. Over 40% of teen users engage monthly with chatbots for emotional support, so your outreach must feel familiar while gently redirecting to clinical resources.

Safeguards should be woven into every step:

• Human-in-the-loop review: Route all AI-generated messages through a clinical team for approval before sending. Experts warn that AI responses in mental health contexts lack consensus on safety, so validation by a human is non-negotiable.
• Transparent disclosure: Include a footer in all AI-generated outreach: *“This message was drafted with AI assistance. Need to talk? Our team is here.”* This mirrors the research’s finding that teens hide AI use—your outreach can normalize openness.
• Limit AI to administrative and preparatory roles—scheduling check-ins, sharing coping tips, or listing local support groups. Avoid any clinical advice, symptom tracking, or crisis intervention language.
• Use developmentally appropriate tone in all AI content, avoiding jargon or clinical terms. Teens prefer non-judgmental, immediate responses—qualities that make AI appealing but also require careful curation to avoid harmful validation.
• Monitor engagement closely and adjust strategies quarterly. The research notes that AI’s helpfulness doesn’t equal efficacy, so track whether outreach actually drives teens toward professional care.

For centers partnering with AI Business Sites, the platform’s AI content engine can generate drafts of outreach emails or social posts grounded in your clinical guidelines, while your team retains final approval. This approach lets you scale content without sacrificing safety—turning AI’s reach into a reliable ally in early intervention.