Therapy Without Humans? The Dangers of AI-Driven Mental Health Care

A recent paper asks "Should AI Replace Therapists?"

Many professionals are anxious about AI’s impact on jobs, from graphic design and computer science to bookkeeping and customer service. Mental health is no exception.

The notion that an algorithm could replicate a therapist’s role is unsettling. After all, the core of therapy isn’t just techniques or content: it’s the human relationship. The therapeutic alliance, trust, and empathy aren’t optional. Instead, these elements are the foundation of effective care.

At PsychVox, we generally focus on applied mental health research, whether that’s culturally-sensitive care, reducing self-harm, or supporting new mothers through postpartum depression. But today, we’re exploring a different kind of study. Published at The Association for Computing Machinery Conference on Fairness, Accountability, and Transparency, it tackles a difficult question: should AI replace therapists?

What They Found: Two Big Failures

In the study, researchers from Stanford, Carnegie Mellon, and the University of Texas evaluated leading AI models, including ChatGPT’s latest version, for their potential role as therapists. Grounding their work in established clinical standards, the team reviewed professional guidelines from major medical organizations to define what constitutes “good therapy.”

This included core themes such as the importance of the therapeutic alliance, emotional intelligence, client-centered and empathetic communication, trustworthiness, offering hope, demonstrating genuine interest, and adhering to professional and ethical standards.