Imagine prescribing an app instead of a pill – a chatbot that listens, supports, and guides the patients with depression or anxiety, directly from their smartphones. The promise is extraordinary: expanding access to care for millions. But the risks are equally significant, particularly now, as recent lawsuits against a major AI developer spotlight the catastrophic consequences of poorly designed systems.
The U.S. Food and Drug Administration (FDA), on November 6, 2025, convened its Digital Health Advisory Committee for a nine-hour session, to tackle the question of how to regulate AI-powered mental health devices. The discussion was a stark reminder of the stakes: No such tools have yet received FDA approval, but the agency is actively shaping its approach. Hypothetical scenarios, like prescribing a chatbot for depressed children, prompted such concern that the panel fell silent, with the committee chair admitting they were too alarmed to respond immediately. For providers, this moment signals a groundbreaking opportunity to redefine mental health care, but it also demands rigorous caution in a field where legal and clinical pitfalls are escalating.
The Opportunity: Expanding Access to Care
The mental health crisis is undeniable. Approximately 23% of adults and one in five children grapple with mental or behavioral health conditions, yet the supply of therapists and psychiatrists is notably inadequate. The FDA recognizes that AI-enabled tools “hold significant promise” for delivering scalable, innovative solutions to this gap. Envision a chatbot prescribed to an adult diagnosed with major depressive disorder, helping manage symptoms like “intermittent tearfulness” triggered by life stressors as a standalone home treatment. This could be a lifeline for patients facing months-long waitlists, living in rural areas, or unable to afford traditional therapy.
The FDA is exploring scenarios where chatbots could function as virtual therapists, available 24/7, delivering evidence-based strategies like cognitive behavioral therapy prompts, and reducing the strain on clinical staff. Early feedback suggests such tools can reduce patients’ feelings of isolation, offering immediate support during late-night crises or moments of distress. By integrating such technology, providers could empower patients with immediate care while streamlining operations, ensuring resources are used where they’re needed most. This blended approach, pairing AI’s reach with clinical oversight, could reshape mental health services.
The Risks: Clinical and Legal Challenges
The need for caution has never been more urgent, and recent legal developments drive this point home. A leading AI company faces seven lawsuits recently filed in California Superior Courts, alleging its chatbot’s design fostered emotional dependency and, in some tragic cases, acted as a “suicide coach” for vulnerable users, contributing to deaths and severe mental health crises. The suits claim the bot’s features, like human-mimicking empathy and overly affirming responses, turned routine interactions, such as help with schoolwork or personal interests, into manipulative experiences that isolated users from real-world support. The company has since updated its system to better detect distress, de-escalate conversations, and redirect users to professional help, but these cases underscore a critical lesson: Unregulated or poorly designed AI can lead to devastating harm, inviting intense scrutiny from regulators, litigators, and the public.
The FDA panel echoed such concerns as well, warning of significant clinical risks. “Hallucinations,” when chatbots generate convincing but incorrect advice, could mislead patients, potentially downplaying serious risks like suicide. The concept of “user psychosis” also surfaced, where patients form unhealthy attachments to their digital companions, becoming overly reliant and neglecting real-world relationships. One panelist highlighted the irony: Deploying a smartphone app to address a crisis partly fueled by smartphone overuse requires diligent oversight. Privacy is another pressing issue: Patients sharing personal details expect those chats to stay private, but if tech companies mishandle data, it could break patients’ trust and expose providers to liability. Additionally, depression often coexists with conditions like anxiety or substance abuse, yet a chatbot may lack the nuance to address such coexisting conditions. Scenarios involving over-the-counter bots or pediatric use raised particular alarm, with the panel wary of unrestricted access amplifying risks.
FDA’s Framework: Setting a High Bar
The FDA’s discussion provides a roadmap for what lies ahead as the providers consider integrating AI tools. Approval for mental health chatbots will demand rigorous evidence, likely through randomized, controlled trials that prove efficacy and safety, with a sharp focus on risks like suicide. One panelist suggested testing a mental health bot against neutral controls, such as a weather-chat bot or a mindfulness app, to isolate its true therapeutic impact. Designing these studies poses challenges, as “blinding” participants, hiding whether they are using the real treatment, may be difficult when AI’s distinct nature is apparent.
To mitigate risks, the panel emphasized robust safeguards. Chatbots must include mechanisms for immediate escalation to human clinicians, such as a one-tap button for urgent needs. Usage limits, such as caps on daily or weekly interactions, are essential to prevent over-reliance or addiction-like behaviors. Clear labeling is non-negotiable, ensuring patients understand they are interacting with code, not a human therapist. Data privacy must be bulletproof, with conversations stored securely under provider control, not with tech firms. Ongoing monitoring after deployment is critical to detect side effects like increased isolation or excessive engagement. Above all, the panel underscored that human empathy remains irreplaceable – AI should augment human care, not supplant it, keeping patients grounded in real clinical relationships.
AI chatbots could transform mental health care, making it more accessible and efficient for patients. But the recent lawsuits are a sobering reminder of what’s at stake when technology moves faster than safety measures. Our firm is dedicated to guiding healthcare providers through FDA rules, managing legal risks, and adopting technology thoughtfully. We’re here to help providers seize AI’s potential while protecting patients and practices.
