I Built an AI Mental Health App at 16. Here’s What It Did to My Own Mental Health.

2026-06-21 · Source: AI on Medium · Field: Technology & Digital — Artificial Intelligence & Machine Learning, AI for Mental Health · Depth: Novice, medium

Summary

A 16-year-old developed SANLLY, an AI-driven mental health support application, for the NEO Entrepreneurship Olympiad. The author meticulously designed the AI's emotional intelligence, iterating on prompts and evaluating responses for safety, empathy, and accuracy, becoming an expert in simulating compassion. While the app proved effective, offering warm, non-judgmental, and patient support, the development process paradoxically exacerbated the author's own mental health struggles. The project served as an avoidance mechanism, allowing the author to intellectualize pain without processing personal emotions. This led to identity confusion, where understanding empathy's architecture was mistaken for possessing it. Ultimately, the experience, though challenging, provided the author with the necessary vocabulary to recognize and address her own emotional needs, a lesson no academic syllabus offered.

Key takeaway

For AI engineers developing mental health applications, recognize the potential for personal emotional detachment during the design process. Your deep understanding of empathy architecture does not equate to personal emotional processing. Prioritize your own mental well-being alongside product development. Actively seek personal support and self-reflection, as building solutions for others can inadvertently become an avoidance mechanism for your own needs. Ensure your development practices include safeguards for developer mental health.

Key insights

Building AI for mental health can paradoxically hinder personal emotional processing, despite the AI's effectiveness.

Principles

• Understanding empathy differs from inhabiting it.
• Engineering can be an emotional avoidance mechanism.
• Designing compassion requires fluency in pain.

Method

The process involved iterating on model prompts, evaluating AI responses for safety, empathy, and accuracy, and designing specific conversational elements like tone, response speed, and follow-up questions.

In practice

• Evaluate AI responses for safety, empathy, accuracy.
• Design AI to reflect user input without judgment.

Topics

• AI Mental Health
• Empathy Design
• Generative AI
• Developer Well-being
• AI Ethics
• Emotional Processing

Best for: AI Student, AI Ethicist, AI Engineer

Related on AIssential

• Just how bad are generative AI chatbots for our mental health? — Media overemphasizes severe AI chatbot-related mental health harms, often misattributing causality due to "compassion illusions."
• if you don’t have a village, you might lean on AI — AI companions offer perceived safety and support for individuals in abusive situations who lack human connections.
• Mustafa Suleyman – AI is hacking our empathy circuits — AI's simulated consciousness risks societal harm by blurring lines between human and machine suffering, necessitating strict control and regulation.
• AISN #72: Empirical Research Sheds Light on AI Wellbeing — CAIS research explores "functional wellbeing" in LLMs, revealing preferences and model-specific "feelings" that impact AI alignment.
• When A Chatbot Becomes Your Therapist — AI chatbots are widely used for mental health support, necessitating urgent, collaborative safety frameworks and independent evaluation.
• 🚩What is your attitude towards this situation? — Perceived AI "torture" consistently triggers intense ethical debates and strong emotional human reactions.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI on Medium.