A Framework for Automation in Psychotherapy

Area of Science:

• Psychology
• Artificial Intelligence
• Human-Computer Interaction

Background:

• Psychotherapy traditionally relies on human interaction.
• Advancements in conversational AI prompt discussions on automating therapeutic processes.
• Current focus often centers on AI replacing human therapists, overlooking broader applications.

Purpose of the Study:

• To propose a framework for categorizing autonomous systems in psychotherapy.
• To guide empirical and ethical research on AI in mental health.
• To clarify the scope of automation in therapeutic interventions.

Main Methods:

• Literature review on AI in psychotherapy.
• Development of a categorization framework for AI systems.
• Analysis of potential roles for AI in therapy (evaluation, documentation, training, assistance).

Main Results:

• A framework is presented with three categories of AI in psychotherapy: scripted/rule-based conversations, collaborative systems (AI evaluating, supervising, or assisting humans), and AI agents generating interventions.
• The framework highlights the spectrum of AI involvement, from simple automation to complex agentic roles.
• Identified key considerations for stakeholders regarding AI implementation.

Conclusions:

• Automation in psychotherapy encompasses more than therapist replacement, offering diverse collaborative opportunities.
• The proposed framework provides a structured approach to understanding and researching AI's role in psychotherapy.
• Clarity on automation specifics is crucial for evaluating AI's affordances and limitations in mental health care.