AI and Psychotherapy: Fantasies, Resistances, and Responsible Clinical Uses

Written by Jennifer Elalouf | Jan 14, 2026 8:50:58 AM

Why Artificial Intelligence Disrupts Psychotherapeutic Practice

Artificial Intelligence (AI) is gradually asserting itself in the field of mental health. Writing assistance tools, augmented clinical platforms, conversational agents, data analysis: the uses are multiplying, often faster than the frameworks of thought that allow for their integration.

For many psychotherapists, this emergence elicits contrasting reactions: curiosity, caution, skepticism, sometimes rejection. These positions are often caricatured as a divide between modernity and conservatism. In reality, they reflect profound clinical, ethical, and psychological issues that must be considered rather than dismissed.

This article offers a nuanced reading: understanding the fantasies and resistances related to AI, identifying what it cannot do, and outlining the conditions for responsible clinical use, compatible with a demanding psychotherapeutic practice.

Fantasies Surrounding AI: What We Project onto Technology

AI as a Figure of Total Knowledge

One of the most common fantasies is that of an omniscient AI, capable of understanding, predicting, and interpreting better than humans. This representation fuels the fear of a loss of clinical freedom, even leading to a normalization of practices.

However, AI systems remain fundamentally dependent on data, frameworks, and goals defined by humans. They possess neither clinical intuition, nor a lived understanding of subjectivity, nor access to the symbolic dimension of psychic experience (Mandal et al., 2025).

AI as a Threat of Dehumanization

Another major fantasy associates AI with the disappearance of human relationships. From this perspective, any technological mediation would be experienced as a depletion of the therapeutic bond.

This fantasy deserves to be examined: clinical practice has never been “purely” relational. It has always relied on mediations—writing, institutional frameworks, diagnostic tools—without this negating the encounter. Thus, the question is not the presence of a tool, but the place assigned to it.

AI as an Unsettling Double of the Therapist

Finally, AI can be perceived as a figure of replacement: faster, more available, potentially less costly. This representation directly touches upon professional identity and the subjective value of clinical work.

Current research shows, however, that AI cannot support a complex therapeutic process, nor work with the ambivalence, temporality, and uncertainty inherent in psychotherapy (Sharma et al., 2022).

Resistances from Therapists: A Clinical Rather Than Ideological Reading

Resistances to AI do not stem from a rejection of progress. They can be understood as clinical signals.

Defending the Framework

The therapeutic framework relies on stability, confidentiality, and reliability. Any new technology legitimately questions:

  • data security,
  • respect for professional secrecy,
  • continuity of the framework.

These concerns are elaborated in a more operational manner in the article "Confidentiality, GDPR, and Digital Tools for Therapists: What You Really Need to Know."

Protecting the Space for Clinical Thought

Another common resistance concerns the fear of a standardization of clinical reasoning. Many therapists fear that digital tools may impose categories, grids, or logics that are foreign to their way of understanding the patient.

This vigilance is well-founded: poorly designed technology can impoverish clinical practice. However, technology conceived as a reflective support can, conversely, sustain the continuity and depth of the work.

What AI Can Actually Bring to Psychotherapeutic Practice

Organizational and Cognitive Support

The most significant contributions of AI today pertain to:

  • aiding in the structuring of clinical notes,
  • synthesizing complex information,
  • long-term monitoring of situations,
  • reducing administrative burdens.

These indirect uses have a real clinical impact: they free up psychic space for the session itself.

Enhancing Clinical Continuity

Some studies show that well-integrated digital tools can improve the continuity of follow-up and the coherence of care without substituting for the therapeutic relationship (Li et al., 2023).

AI then becomes a discreet assistant, serving the clinician, rather than an actor in the therapeutic process.

What AI Cannot Do (and Will Not Do)

It is essential to set clear limits.

AI cannot:

  • experience subjective presence,
  • tolerate ambivalence and uncertainty,
  • work with what is left unsaid and silences,
  • engage in a singular psychic temporality,
  • assume clinical responsibility.

These dimensions constitute the core of psychotherapeutic work and remain irreducibly human.

Towards Responsible Clinical Uses of AI

A responsible use of AI in psychotherapy is based on several principles:

  • AI never replaces the therapist,
  • clinical judgment remains central,
  • data is strictly protected,
  • tools are chosen to support thought, not to normalize it,
  • transparency towards the patient is guaranteed.

In this perspective, some professional platforms have chosen AI designed as a tool for reflective support, integrated within a strict ethical framework that respects confidentiality and clinical work.

AI, Ethics, and Sustainability of Practice

One often-overlooked point concerns the sustainability of therapeutic practice. By supporting organization, clinical memory, and administrative burden, AI can indirectly contribute to preventing professional burnout.

These issues are discussed in “Preventing Emotional Burnout in Therapists: Recognizing, Understanding, and Acting.”

Conclusion: Thinking About AI Rather Than Suffering Its Consequences

AI acts as a revealer: it brings to light our representations of care, knowledge, and the therapeutic relationship. The fantasies and resistances it evokes are not obstacles to be eliminated but clinical materials to consider.

When integrated with discernment, AI can become a valuable support, allowing psychotherapists to preserve what lies at the heart of their profession: presence, listening, and clinical thought.

References

Clavier, B., & Botbol, M. (2023). Repenser la prise en charge de la santé mentale à l’ère de l’intelligence artificielle. L’Information psychiatrique, 99(4), 291–298.

Li, H., Zhang, R., Lee, Y. C., et al. (2023). Systematic review and meta-analysis of AI-based conversational agents for promoting mental health and well-being. NPJ Digital Medicine, 6, 227.

Mandal, A., Chakraborty, T., & Gurevych, I. (2025). Towards privacy-aware mental health AI models: Advances, challenges, and opportunities. ACM Computing Surveys.

Sharma, A., Lin, I. W., Miner, A. S., Atkins, D. C., & Althoff, T. (2022). Human–AI collaboration enables more empathic conversations in text-based peer-to-peer mental health support. Proceedings of the ACM on Human-Computer Interaction, 6(CSCW2).