AI Agents Are Coming: 5-Stage Taxonomy of Language-Based AI Systems for Psychiatry, Psychotherapy, and Counseling

Clinical Snapshot

Key Findings

• P Value: Not reported — no statistical testing conducted

• Effect Size: Not applicable — no quantitative effect size reported; this is a theoretical framework paper

• Primary Outcome: A 5-stage taxonomy for language-based autonomous systems in mental health: Level 1 (Knowledge — static benchmark tasks), Level 2 (Elementary — dynamic therapeutic microskills), Level 3 (Integration — cross-module consistency, basic case conceptualisation, blended therapy under human oversight), Level 4 (Saturation — therapist-in-the-loop, minimal supervision), Level 5 (Mastery — technically capable of autonomous therapy)

• Nnt Or Sensitivity: Not applicable — no clinical trial data; the key conceptual finding is that Level 4 or 5 technical performance does not automatically confer full treatment effectiveness, even with high treatment fidelity

• Confidence Interval: Not reported — no empirical data presented

Clinical Application

The taxonomy is immediately applicable as a conceptual reference framework for clinical governance discussions, ethics committee reviews, and research protocol design. However, operationalisation for regulatory or procurement purposes would require further empirical validation and consensus development. Blended therapy models (Level 3) are the most immediately feasible clinical application given current technology maturity. In Australia, the TGA regulates software as a medical device (SaMD) under the Therapeutic Goods (Medical Devices) Regulations 2002, and language-based autonomous mental health systems at Levels 3–5 would likely require TGA classification and conformity assessment. The RACGP and RANZCP have not yet published specific guidelines on autonomous AI in psychotherapy, making this taxonomy potentially valuable for informing future position statements. The PBS does not currently fund AI-delivered psychotherapy, and Medicare Benefits Schedule items for psychological therapies (e.g., Better Access initiative) are restricted to registered practitioners — autonomous systems at any level would require legislative amendment for reimbursement. The Australian Digital Health Agency's National Digital Health Strategy 2023–2028 provides a relevant policy context. Headspace, Beyond Blue, and other digital mental health platforms operating in Australia would benefit from a validated version of this taxonomy for service governance. Equity considerations are particularly salient in Australia given the mental health workforce shortage in rural and remote areas, where autonomous systems might be considered for access expansion. Mental health clinicians, psychiatrists, psychologists, counsellors, and digital health developers evaluating or deploying language-based autonomous systems; health service administrators and regulators developing governance frameworks for AI-assisted mental health care

Abstract

The rapid evolution of large language models has accelerated the development of agentic artificial intelligence (AI) systems capable of pursuing autonomous goals, creating an urgent need for structural frameworks in psychiatry and psychotherapy. While existing classifications often draw parallels to autonomous driving, this paper argues that the mental health domain requires a distinct, domain-specific theoretical foundation, as the 2 domains differ fundamentally in their semantic, ideographic, and epistemological demands. Furthermore, they differ in their end goals, for which we introduce terms such as agentic guidance capability. To guide clinicians and researchers through these developments, we propose a 5-stage taxonomy for language-based AI systems that differentiates technical functionality from clinical effectiveness. The taxonomy progresses from level 1 (knowledge level), in which systems perform static benchmark tasks, to level 2 (elementary level), characterized by dynamic engagement in specific therapeutic microskills. At level 3 (integration level), systems achieve consistency across and within modules, as well as basic case-level conceptualization suitable for blended therapy under human oversight. Level 4 (saturation level) describes therapist-in-the-loop systems capable of autonomous functioning with minimal supervision, whereas level 5 (mastery level) represents AI systems that are technically capable of performing autonomous therapy. By distinguishing technical functionality from clinical effectiveness, we conclude that level 4 or level 5 performance does not automatically translate into full treatment effectiveness, even if high treatment fidelity can be achieved. We conclude by emphasizing the need to shift benchmarking from static knowledge tests to dynamic evaluations of therapeutic capabilities in order to safely navigate the transition toward autonomous care.

References

1. 1.Schuster, R., Plessen, C. Y., Carlbring, P., & Walther, A. (2026). AI Agents Are Coming: 5-Stage Taxonomy of Language-Based AI Systems for Psychiatry, Psychotherapy, and Counseling. JMIR Mental Health. https://doi.org/10.2196/91746