As artificial intelligence (AI) becomes more integrated into mental health care, it raises important questions around ethics, boundaries, and emotional safety. While AI tools offer exciting possibilities—like supporting therapists with note-taking, improving diagnosis, or increasing access to care—they also come with serious challenges that can’t be ignored.
One of the biggest concerns is privacy. AI tools often rely on large datasets, which may include sensitive personal information. How that data is stored, shared, and used must follow strict confidentiality rules, just like traditional therapy. Without strong data protections, clients’ emotional safety and trust are at risk. Those using AI to have conversations about their mental health may benefit from engaging with a tool that provides information, or interpretation of interpersonal situations. Privacy and data collection of health and personal information are concerns, especially for those who are vulnerable and seeking help. And, it's worth noting that many of those using AI for mental health concerns, are coming to therapy and reflecting on the same situation within a trusting relationship with a trained professional. In this regard, I wonder if AI is something that can create dependence and reinforce helplessness with its 24/7 access, and immediate responses, rather than foster greater connection to self and inner wisdom?
Another challenge is accuracy and bias. AI can reflect the same systemic biases found in the data it was trained on. That means without careful oversight, it might misinterpret emotional cues, offer inappropriate advice, or reinforce harmful stereotypes—especially for marginalized groups.
Educators, therapists, and clients need clear guidelines for how AI is used in therapeutic spaces. Open conversations around consent, boundaries, and responsible use are essential for protecting emotional well-being. If your therapist is using AI for their notes, this should be disclosed and you should have provided consent.
Ethically, AI is not a therapist. It cannot replace the nuanced, human connection that licensed professionals offer. It’s important for users to understand the limits of AI support and not to mistake digital tools for clinical care. As the field evolves, the priority must remain the same: centering human dignity, empathy, and safety in every interaction—digital or otherwise. AI can assist in healing work, but it should never replace the heart of it.