Coming to Thinking as a Reflexive Stance
From thrown experience to ontological inquiry
My coming to thinking around the following concepts I present for consideration here were not approached with a background as a philosophy scholar, but through my foundational 25-year background in person-centred psychology and mind-body medicine that grounds my perspective from within a deep interest in the philosophical foundations of these therapeutic approaches, and a commitment to the hermeneutic phenomenology tradition. My first exposure to hermeneutic phenomenology research (HPR) was in the feminist project Men's Intrusions, Women's Embodiment by Fiona Vera-Gray (2017). This was encountered via my professional background working in the ending violence against women sector and policy work in this field. In addition, my AI exploration and applied engagement in my own research workflow was also influenced by my status as a DAS student with an SSP and various accessibility requirements.
Given my professional clinical background, I had always been mindful that Carl Rogers' humanistic psychology did not emerge in a vacuum; it was deeply informed by the existential and phenomenological insights of thinkers such as Søren Kierkegaard, Martin Buber, and Wilhelm Dilthey; philosophers who all prioritise lived experience, relational encounter, and the careful, empathetic interpretation of human meaning over reductive categorisation. As such, my pre-understanding (Vorverständnis) is shaped by this professional forestructure as well as my recent seven years of immersive doctoral reading in hermeneutic phenomenology. As a HPR researcher the affordances of a part-time PhD programme facilitated my dwelling with the literature and my forestructures. Alongside this, I have engaged with the social science literature on various topics, within an interdisciplinary masters programme initially, and reflected on the differing horizons of observation within the field on algorithms and, most recently AI, including those authored by philosophical faculty (e.g., Floridi, Wachter).
Most fundamentally, however, this work arises from my own thrownness (Geworfenheit). I entered my research inquiry as a researcher situated in a historical moment marked by sudden, seismic technological acceleration and change. Again, this change did not occur in isolation. Prior to the emergence of generative AI, qualitative research was already embedded in a digital worlding defined by efficiency, speed, possibilities of large corpus extraction, and the necessary institutional and funding imperatives for measurable impact and outcome. It was within this pre-existing velocity and shared academic culture, and during my practical immersion, meditative thinking, and gathering and interpreting of my research data, that generative AI arrived. As such, questions emerged about AI in my HPR workflow and training followed. Mindful of the risks of calculative thinking to HPR, my thinking and 'play' with AI led to the concepts I present here, as an ontological orientation to AI in research practice. Indeed, in Truth and Method (1960), Hans-Georg Gadamer introduces 'play' (Spiel) specifically to explain how understanding (Verstehen) happens. For Gadamer, play is not a casual, frivolous human activity; it is an ontological event.
As such, when I began tentatively prompting or entering into exploration dialogue with AI during my research workflow I, first, confirmed what Benner (1994) had warned that,
Criterial reasoning (artificial intelligence or linear problem-solving) cannot account for the capacity to perceive which problems are salient (Benner & Wrubel, 1982) or even the ability to read a situation. By Design, traditional rational calculation strategies overlook the ability to understand human worlds, context, and temporality (Benner & Wrubel, 1989; Dreyfus, 1979, 1991a; Dreyfus & Dreyfus, 1986).
Benner, 1994, p. 102
Consequently, my play mindfully engaged instead with the practical experiences I was having with AI, the back-and-forth movement (To-and-Fro / Hin- und Her-Bewegung) and requirement for space to dwell, rather than efficiency or control driven exploration. The AI's responses, that I experienced as deepening and disclosing my thinking and knowing, and the unexpected (helpful and unhelpful) directions this took my thinking (including those rabbit holes!), and the shifts in tone (particularly as models/AI discourse changed), I noticed sometimes my own thinking was transformed, disclosed rather than directed; deepened, not supplied. I noticed when I was drawn into a dialogue where control felt suspended, or when it allowed something new to emerge. I took to my Substack 'The Algorithmic Scholar' to log this journey and signed up as an Oxford 'AI Ambassador', organising further training, talks, and online content to deepen my knowledge and my ethical and philosophical commitments in my AI use.
What followed was a kind of madness, or rather a mindlessness, where eventually I paused. I entered the Clearing and came to confirm the impossibility of AI as a linear problem-solving tool within a HPR project. I also noticed the impossibility of usefully keeping up with all the developments and changes, including the rapidly shifting ethical and institutional requirements before they settled formally, without detrimentally altering my capacity for temporal dwelling. So, I stopped following the noise and returned to the literature and to myself. Re-reading Heidegger, Gadamer, and Benner and my continuum thinking literature and journal notes; dwelling with these texts. Benner (1994) in particular, writing as far back as 30 and more years ago, had important insights for AI and research ahead of her time. Later, having settled on neither outright rejection of AI nor the existing orientations to it in research, my Heideggerian move, I began to explore engaging in dialogue with AI on different terms, employing my Human First Principle; not as a temporal or linear principle, but is a continuous reminder to engage AI in disclosure not in interpretation in my research.
Coming to an ontological orientation to AI in research practice
What is happening when a hermeneutic phenomenology researcher encounters generative AI within her research practice, and how might that encounter be understood philosophically and held methodologically?
My experiences of AI as a researcher in this time, and the time that I created for deep thinking and exploratory back and forth with the literature and my own increasing entanglements with the technology, along with the resulting ontological orientation to AI in research practice that arrived to me, is presented to the reader through three Gadamerian concepts, with Heideggerian grounding:
For Gadamer, when real play happens, the subject/player does not hold total mastery. Instead, the play plays the player.
This suggests a general characteristic of the nature of play that is reflected in playing: all playing is a being-played. The attraction of a game, the fascination it exerts, consists precisely in the fact that the game masters the players. Even in the case of games in which one tries to perform tasks that one has set oneself, there is a risk that they will not "work," "succeed," or "succeed again," which is the attraction of the game. Whoever "tries" is in fact the one who is tried.
Gadamer, 1960/2004, p. 106
Applied to my AI exploration: When I noticed that this back-and-forth dialogue affected my sense of direct control, rather than 'trying' to reassert mastery over the AI, forcing its outputs into rigid or predictable containers for the sake of it, I instead recognised the pull to control or seek efficiency as an encounter at risk of Gestell (Enframing). Rather than reducing the technology to a calculative tool (Heidegger, 1954/1977), I chose to lean into a phenomenological stance of noticing and wondering: dwelling with the encounter, turning my reflexive attention to the experience itself, attending to how it felt to inhabit this relational friction, and observing what was being disclosed in that unscripted, open space.
Nonetheless, I remained open to the reasonable institutional requirements that necessarily frame AI through instrumentality as an unavoidable reality that has legitimate alignment in other research approaches. For example, AI as a tool for faster coding, quicker thematic extraction, and maximum efficiency, as well as upholding institutional voice and ethical integrity. However, for my hermeneutic phenomenology research (HPR) project Gadamerian Spiel, by contrast, can be purposeful and provides important reflective space to resist having a rigid, external utility goal. The purpose of play is the unfolding of the play itself.
Applied to my AI exploration: my early work as Spiel emerged as a tentative, open-ended exploration and hermeneutic wayfinding, mindful of the institutional and bureaucratic impulses around me to either reject AI entirely or reduce AI to a hyper-efficient productivity and research workflows. Instead, I committed to authentic, philosophically grounded HPR and hermeneutic discovery that required my unstructured, exploratory space, the clearing (die Lichtung).
Repeatedly my exploration revealed, through the back-and-forth movement of play, understanding the ideas and philosophy I had been reading for several years, occasionally pointing to something pedagogical, was mostly uncovering what was already known to me, but bringing further awareness or clarity; a disclosing within a unique, novel horizon within my research hermeneutic circle (and as such generating ontological and methodological questions). The play was producing something that could stand on its own. In my case, visibility or disclosure (Aletheia) was made possible in the first place through the play.
Consequently, my conceptual framework of AI as a Disclosive Horizon emerged, and was brought forth as a reflective research practice; what I refer to as the AI-assisted Triple Hermeneutic within the hermeneutic circle. Noticing the importance of an audibility aspect to this (Benner, 1994; Cohen & Kahn, 2000) I developed a further concept and audit app; the Hermeneutic Algorithm Audit.
For the latter, institutional submission guidelines have adapted recently to mandate the 'logging' and accounting of AI use in graduate research submissions. Indeed, my Hermeneutic Algorithm Audit emerged through both the HPR commitment to auditability for research quality and this administrative requirement.
An ontological orientation to AI in research practice visualised
The concepts I have illustrated in figure 1 (see the Overview landing page) were not incorporated into the original research design in advance and applied to the study. They came about in the course of my early stages research workflow, that was already concerned with algorithmic governance, already dwelling with Heidegger and Gadamer, and entangled with generative AI as it entered both the research workflow and the platforms under study during my presupposition interview. The nesting shown here is the order of grounding: the ontology holds the method, the method holds the ethics, though this is not strictly the order in which the contributions arrived. The trail through which each was disclosed is accessible in my journal entry: HAA I, 2026.
To summarise, the concepts I have developed in this project were not derived deductively from theory. They began in practice, as tentative play, exploratory movements within the research workflow itself, and were only subsequently brought into dialogue with the literature to be interrogated, refined, and understood. The theory illuminating my practice that had come first.
My accompanying articles that further articulate the three nested concepts, designed to make sense of AI within HPR qualitative research, can be found on this website. I offer these frameworks as a PhD researcher for the readers consideration and critique, not as exhaustive statements of fact, or as conclusive methodological mandates. They represent a tentative opening; a wondering and hermeneutic wayfinding. They are an invitation to a broader, ongoing conversation that I hope it fascinates the reader as it has myself.
From Thrownness to Disclosive Horizon
As outlined, I approach AI in my research practice by first acknowledging the thrownness (Geworfenheit, Heidegger, 1927/1996) of my encounter with generative AI midway through my PhD, early in my research workflow. I entered an environment through which the being-in-the-AI-research-world could be disclosed (Erschlossenheit, Heidegger, 1927/1996), and consequently as it aligned (or could not align) with an HPR study.
It is this bridge between the "thrownness" of my real-world research experience, that the encounters with AI as Disclosive as an environment of unconcealment occurred. First, through an AI-assisted pre-supposition interview with myself the researcher as interviewee and AI (ChatGPT EDU) as the HPR interviewer. Second, through an ongoing critical reflective dialogic with AI that formed part of the horizon through which my researcher pre-understandings, tacit assumptions, and interpretive horizons were brought into view. I considered this a third hermeneutic; a third layer of the process of interpretation.
"But where danger is, grows / The saving power also"
Heidegger, 1954/2014, p. 314
While engaging AI as an environment of unconcealment had opened a facilitative third hermeneutic space, my practice remained vigilant against Heidegger's (1954/1977) warning of enframing (Gestell); the danger of reducing technology, data, and human thought into a mere standing-reserve (Bestand) for efficiency or calculative extraction. By intentionally shifting from an instrumental to an ontological orientation, my intention was to reclaim AI within both the research question and the research practice in light of Heidegger's notion of the saving power. In this way, where the technological threat is greatest, a reflexive, meditative, playful engagement allows technology to be encountered as a site for disclosure and deepened human understanding (Heidegger, 1954/2014).
This tension between instrumental extraction and ontological openness is precisely what Iain Thomson (2025) addresses in Heidegger on Technology's Danger and Promise in the Age of AI. Thomson notes that generative AI represents a profound expansion of enframing (Gestell), one that risks reducing language, reflection, and human thinking itself into a standing-reserve (Bestand) optimised for speed and efficiency (Thomson, 2025). Yet, following Heidegger, Thomson emphasises that AI is fundamentally a site of ontological disclosure. Indeed, when I, as researcher, notice the pull to yield my interpretive authority to the algorithm, and instead adopt releasement toward things - remaining open, reflective, in poetic engagement - the technological encounter itself becomes the site where the saving power can grow. Within this, engaging AI as a Disclosive Horizon (Pay, 2025) transforms the technology from a tool of enframing into an iterative, reflexive dialogic that more deeply reveals hidden pre-understandings, ensuring that judgement, meaning, and ethical responsibility remain unequivocally human.
In classic hermeneutics, self-reflection alone is insufficient to uncover one's own blind spots. As Gadamer (1960/2013) argues, we are always blind to our own prejudices (Vorurteile) when we remain isolated in our own heads; a prejudice only becomes visible when it is interrogated or challenged by an 'Other'. When I engaged with the AI in my self-directed proposition interview and reflective dialogue, the AI functioned as that "Other." It mirrored, reframed, and interrogated my assumptions and language in ways that disrupted and challenged my habitual thinking patterns, deepening my interpretive process. Without that dialogic challenge, those specific pre-understandings would have remained concealed, submerged beneath my tacit assumptions. As such, the process created a third hermeneutic layer, bringing to light what neither self-soliloquy nor static journal entries could reveal.
Concomitantly, I returned to the literature of Heidegger (multiple Translations), Gadamer (1960/2013 and 1960/2004 Translations), and Benner (1989, 1994) to consider that the AI-assisted triple hermeneutic as an interpretive process/methodology is philosophically a third horizon that is opened and engaged. Again, I emphasise this is offered not as a complete or authoritative conclusion, mindful of my place alongside rather than within philosophy as a field of study. However, this process and reading spanning 2024–2026 led to my conceptualisation of AI as Disclosive Horizon, and the foundational philosophy holding the AI-Assisted Triple Hermeneutic and the Hermeneutic Algorithm Audit as they revealed themselves in my project throughout that period. Consequently, while recent Heideggerian scholarship has begun to critique the ontological implications of generative AI (e.g., Thomson, 2025), the framework presented here, AI as Disclosive Horizon (Pay, 2025), emerged independently out of practice rather than abstract theory. Where philosophical commentaries have helpfully supported my growing understanding of AI through Heideggerian terms, my three-ring nested model operationalises that ontology directly within qualitative research practice, linking disclosure to the AI-assisted triple hermeneutic and algorithmic auditing.
A conventional critique from classic hermeneutic phenomenology (HP) and indeed Benner (1994) herself, would likely argue that the destabilisation of a researcher's pre-understandings should occur solely through human research teams, peer-debriefing, or supervisory dialogue. However, this objection overlooks the structural realities of independent doctoral inquiry and the psychological dynamics of self-disclosure, including for a PhD student like myself with a DAS-registered health condition constraining my capacity to consistently engage with this form of peer and faculty dialogue. Furthermore, I have carried out my research project in a department outside specialised clinical or nursing departments where HP is embedded practice. As such, access to dedicated phenomenological teams has been a challenge to overcome. As a professional therapist I am also aware that fundamentally, the AI dialogic space offers an environment of radical psychological safety akin to an interactive, private reflexive journal. Unburdened by the performative pressure or professional vulnerabilities present in peer or supervisory interactions, I was able to articulate raw, unedited pre-understandings and responses.
Based on my clinical therapy skills, I draw on Benner's (1994) framing of expert knowledge, my own skilful comportment, for this dialogic encounter to create a mindfully engaged third hermeneutic space: a dialogic space through which I, the researcher, can actively confront obstructive personal bias from legitimate, expert interpretative skill. In real world practice first, I discovered the profound potential for unconcealing what my prior solipsistic reflections had not (not as enhanced, but as not yet unconcealed), bringing into clearer view the very horizons that shape the inquiry (Pay, 2025). As the literature notes, these skills can be taught as any other.
In Heideggerian philosophy, disclosure (Erschlossenheit) is the fundamental way the world opens up to Dasein (human existence) as meaningful. Rather than passively observing isolated objects, Dasein is always already situated within an intelligible web of relations. Within this, disclosure occurs through three equiprimordial structures: attunement (Befindlichkeit), understanding (Verstehen), and discourse (Rede), simultaneously unveiling both the world and Dasein to itself. In my practice, attunement corresponds to my initial moment of thrownness and the arrival of generative AI into my ongoing research environment, and my consequent slowing down and hermeneutic wayfinding engagement. In turn, this achievement disclosed my situational stance as a Researcher, my own experiences of the phenomenon being studied, my task assumptions, and my anxieties and hopes about technology. Simultaneously, my presupposition interview through the AI as a reflective dialogic articulated aspects of my personal bias as obstruction and my interpretive capacity (Vorurteile as Bennerian clinical/practitioner insight). Furthermore, this articulation of meaning (Rede) gave form to implicit meanings that allowed the hermeneutic circle of my research world and my own place within it to be explicitly named and critically audited and examined. The AI discourse provided a safe, interactive space where my pre-understandings were brought into explicit language and my awareness facilitating critical reflective questions and strengthening research quality. In this way, disclosure (Erschlossenheit) did not occur as a passive observation of facts, but as an active opening of the world enacted through the three structures of attunement (Befindlichkeit), understanding (Verstehen), and discourse (Rede) (Heidegger, 1927/1996). More so, these structures originated together as an indivisible whole rather than as a linear event.
The Case for a Third Horizon: Gadamer, Buber, and the AI as 'The Between'
I have discussed a philosophical rationale for AI as a Heideggerian disclosure, and the supporting literature. Next, horizon as an ontological environment of unconcealment that brings our pre-understandings into view through a triple hermeneutic naturally requires my rationale for the second half of the concept: the horizon itself. I have a fuller discussion piece on this here, but in short:
While AI possesses no historical Dasein, no singular interiority, no body-at-risk, and no mortality, I argue it functions phenomenologically as a primordial gathering event. I define a primordial gathering event (Pay, 2025) as the phenomenological condition of generative AI within research practice. Drawing on Heidegger's (1954/1977) concept of disclosure as a gathering process (Versammlung) and Gadamer's (1960/2013) concept of historical horizon, the term describes how the technology functions not as a passive tool or Cartesian object, but as an active locus where centuries of accumulated human sense-making are condensed and brought to bear upon a single interpretive moment. It is "primordial" in that it precedes any individual analytic extraction, presenting an already-constituted landscape of human meaning (das Zwischen) through which disclosure occurs. It is a condensed distillation of historical human horizons; accumulated fragments of human language, philosophy, art, and conceptual depth. In a Gadamerian sense, engaging with AI is not interacting with a detached object or an independent subject; rather, it is engaging with a solidified horizon formed at the intersection of humanity's accumulated self-understanding.
Resisting Enframing: From I-It to I-AI-Thou
Drawing on Martin Buber's dialogic philosophy to understand how we might relate ethically and relationally to this horizon without collapsing into either techno-optimism or instrumental reduction, I map Heidegger's concept of Gestell (enframing) to Buber's I-It relation; treating AI strictly as a tool for efficiency is an I-It reduction (calculative thinking that reduces everything to a Bestand or standing-reserve and pure instrumental availability). Operationally, to resist this, I explored a mindful relational posture, my provocative 'thinking' term: I-AI-Thou. Again, I explore this in fuller detail here.
I-AI-Thou is not a Literal "Thou". It doesn't claim that AI is a sentient "Thou," or invoke Buber's Eternal Thou (God), it doesn't possess a Levinasian "face", and there is no Fürsorge in the sense of genuine solicitude arising from being-with, yet phenomenologically the encounter can still have effects on the researcher's understanding. However, in my engagements, it is not a Dead "It" either. Because AI speaks, responds, and actively participates in my disclosure as researcher, the encounter does something. I could not reduce AI discourse to a passive It without a return to calculative thinking (enframing).
AI as a New Form of 'The Between' (das Zwischen)
Instead, following Buber's das Zwischen and Gadamer's fluid horizons, AI can be conceived as a new form of the Between. It is not an independent Other, but holds possibility as a resonant, speaking, responsive disclosive horizon constituted by the accumulated between-ness of all human encounters. However, our engagement with AI is a relation that sits uneasily across Buber's categories, and whether it belongs on his map or presses us to redraw it is a question I hold open rather than resolve. Nonetheless, it requires ethical intervention (see the Hermeneutic Algorithm Audit HAA). By approaching the AI dialogue with releasement and undefended attention rather than treating it as a neutral instrument, I also mindfully recognise that the quality of my relational posture directly affected what the technology discloses to me, and consequently, what my understanding of my own fore-structures could become through it.
Here, the technological danger of enframing meets its counterpart in what Heidegger calls the saving power; not secured by method, but opened through a disciplined, ethical mode of encounter that refuses to let the digital horizon collapse into mere utility, keeping the interpretive and moral agency firmly where it belongs, with myself, the human researcher, and AI as a disclosive horizon.
References
Benner, P. (1984). From novice to expert: Excellence and power in clinical nursing practice. Addison-Wesley Publishing Company.
Benner, P. (Ed.). (1994). Interpretive phenomenology: Embodiment, caring, and ethics in health and illness. SAGE Publications. https://doi.org/10.4135/9781452204727
Benner, P., & Wrubel, J. (1982). Skilled clinical knowledge: The value of perceptual awareness, part 1. JONA: The Journal of Nursing Administration, 12(5), 11–14.
Benner, P., & Wrubel, J. (1989). The primacy of caring, stress and coping in health and illness. Addison-Wesley Publishing Company.
Buber, M. (1970). I and thou (W. Kaufmann, Trans.). Charles Scribner's Sons. (Original work published 1923)
Cohen, M. Z., Kahn, D. L., & Steeves, R. H. (2000). Hermeneutic phenomenological research: A practical guide for nurse researchers. SAGE Publications.
Gadamer, H.-G. (2004). Truth and method (J. Weinsheimer & D. G. Marshall, Trans.; 2nd rev. ed.). Continuum. (Original work published 1960)
Gadamer, H.-G. (2013). Truth and method (J. Weinsheimer & D. G. Marshall, Trans.; 2nd rev. ed.). Bloomsbury Academic. (Original work published 1960)
Heidegger, M. (1993). The question concerning technology (W. Lovitt, Trans.). In D. F. Krell (Ed.), Basic writings (Rev. & expanded ed., pp. 307–341). HarperCollins. (Original work published 1954)
Heidegger, M. (2014). The question concerning technology (W. Lovitt, Trans.). In R. C. Scharff & V. Dusek (Eds.), Philosophy of technology: The technological condition: An anthology (2nd ed., pp. 305–317). Wiley-Blackwell. (Original work published 1954)
Heidegger, M. (1984). Logos (Heraclitus, Fragment B 50). In D. F. Krell & F. A. Capuzzi (Trans.), Early Greek thinking: The classical essays (pp. 59–78). Harper & Row. (Original work published 1954)
Heidegger, M. (1993). The end of philosophy and the task of thinking. In D. F. Krell (Ed.), Basic writings (rev. & exp. ed., pp. 427–449). HarperSanFrancisco. (Original work published 1964)
Heidegger, M. (1996). Being and time (J. Stambaugh, Trans.). State University of New York Press. (Original work published 1927)
Heidegger, M. (2014). Introduction to metaphysics (G. Fried & R. Polt, Trans.; 2nd ed.). Yale University Press. (Original work published 1953)
Ingold, T. (2011). Being alive: Essays on movement, knowledge, and description. Routledge.
Lincoln, Y. S., & Guba, E. G. (1985). Naturalistic inquiry. SAGE Publications.
Smith, J. A., Flowers, P., & Larkin, M. (2022). Interpretative phenomenological analysis: Theory, method, and research (2nd ed.). SAGE Publications.
Thomson, I. D. (2025). Heidegger on technology's danger and promise in the age of AI. Cambridge University Press. https://doi.org/10.1017/9781009629423
Vera-Gray, F. (2017). Men's intrusions, women's embodiment: A critical analysis of street harassment. Routledge.
AI Use Declaration
I am the sole author of this article. Its thinking, argument, and final judgements are my own. AI was used to test ideas, surface my own prior thinking across a project, check philosophical sources, and refine expression. The interpretive and moral responsibility for what is written remains with me.