1.5 Methodological Approach and Structure of the Thesis

This study is a theoretical and conceptual review of the literature employing a hermeneutic approach. It does not collect its own empirical data, but rather combines theoretical positions and existing empirical findings from several disciplines to form a coherent argument.

The theoretical framework is drawn from the philosophy of dialogue (Buber), sociology (Rosa’s resonance and unavailability theories), phenomenology of the body (Merleau-Ponty, Fuchs, Waldenfels), developmental psychology (Stern) and enactive cognitive science (5E cognition).

The empirical basis is drawn from research into psychotherapy, effective factors and therapeutic alliance (Bordin, Grawe, Wampold), from Gestalt therapy research on consensus and validation (Fogarty), and from clinical research into AI and chatbots.

The findings are not merely listed in sequence, but are interpreted phenomenologically and dialogically and related to the research questions. Where clinical application requires it, legal and regulatory framework sources are also included (including the EU AI Act, the German Psychotherapy Act (PThG) and the GDPR).

The methodological contribution thus lies in the interdisciplinary synthesis, which enables the possibilities and limitations of artificial intelligence to be assessed. The argumentation is deductive in nature. It is derived from three premises and a conclusion:

(1) The therapeutic relationship is a key factor in the effectiveness of psychotherapy.

(2) This relationship presupposes intercorporeal resonance.

(3) AI fails because of this ‘inter-corporeality’, as it lacks a body, an ’I’ and a potential ’you’.

It follows that artificial intelligence is structurally incapable of providing a key factor in psychotherapy: what it is denied is not an aspect of the relationship, but the relationship itself.

In light of this, Gestalt therapy is regarded as a method that fosters this resonance through physicality, presence and the dialogical ‘in-between’. This thesis is discussed in the following chapters.

First, the phenomenon of the therapeutic encounter is defined (Chapters 2 and 3), before this framework is applied to AI (Chapters 4 and 5). Consequently, AI is deliberately avoided in the introductory chapters and is only introduced once the previously developed concept of the encounter is firmly established. The concepts follow a clear sequence, from the foundations (inter-embodiment, unavailability) through the ontological rupture (no body, no ‘I’, no possible ‘you’) to the clinical consequences (confluence, sycophancy). The approach remains consistently phenomenological-clinical throughout.

Chapter 2 sets out the theoretical foundations. Drawing on research into therapeutic alliance and effective factors (Bordin, Grawe, Wampold), the relationship is identified as a central factor in the success of therapy. This is followed by a definition of the place of encounter (Buber’s concept of the ‘in-between’) and its characteristics (Rosa’s concept of resonance), before the concept of ‘inter-corporeality’ is further developed with reference to Merleau-Ponty and Fuchs, supplemented by Stern’s implicit, pre-linguistic knowledge of relationships, and integrated into Fuchs’s stage model of empathy.

Chapter 3 elaborates on Gestalt therapy competencies and thereby defines what constitutes Gestalt therapy as applied relational medicine. The content framework is provided by Gestalt therapy consensus research (Fogarty et al., 2016, 2020) and the phenomenological definition by Fuchs and Schmidsberger (2024). Both approaches exemplify the endeavour to define Gestalt therapy through its relational-bodily approach rather than through technique.

Thomas Fuchs’s 5E model, which is presented in Chapter 2.3 as an ecological paradigm of relational medicine, serves as the organising principle. Chapter 3 examines this paradigm from a clinical perspective: it maps the dimensions of 5E cognition onto three central guiding questions, namely with what (3.1), where (3.2) and how (3.3) encounters take place (see Figure 2), and links each of these questions to a fundamental concept of Gestalt therapy. In this way, the catalogue of competencies drawn up by Fogarty’s consensus panel is integrated into the argumentative logic of this work. Buber’s dialogical approach is incorporated throughout.

Chapter 4 examines the reasons why artificial intelligence cannot structurally fulfil the previously defined relational qualities: without a biological body, neither intersubjectivity nor self-resonance is possible (4.1), and without its own boundary of contact, no encounter takes place, but merely an unopposed adaptation (4.2).

Chapter 5 confronts this ontological demarcation with clinical reality. It examines what AI can empirically achieve as a tool and, on this basis, develops a relational concept of the tool (5.1); it analyses the illusion of a digital therapeutic alliance (5.2); and it identifies the clinical, societal and legal risks that arise when the tool is misunderstood as a counterpart (5.3).

Chapter 6 answers the three research questions on the basis of the preceding chapters, formulates recommendations for practice and concludes with an outlook on an ‘embodied cure’ in the age of artificial intelligence.