Encountering Generative AI: Narrative Self-Formation and Technologies of the Self Among Young Adults

Abstract

This paper examines how young adults integrate generative artificial intelligence chatbots into everyday life and the implications of these engagements for the constitution of selfhood. Whilst existing research on AI-mediated subjectivity has predominantly employed identity frameworks centered on social positioning and role enactment, this study foregrounds selfhood—understood as the organization of subjective experience through narrative coherence, interpretive authority, and practices of self-governance. Drawing upon Paul Ricœur’s theory of narrative self and Michel Foucault’s concept of technologies of the self, the analysis proceeds through in-depth qualitative interviews with sixteen young adults in Norway to investigate how algorithmic systems participate in autobiographical reasoning and self-formative practices. The findings reveal four dialectical tensions structuring participants’ engagements with ChatGPT: between instrumental efficiency and existential unease; between algorithmic scaffolding and relational displacement; between narrative depth and epistemic superficiality; and between agency and deliberative outsourcing. The analysis demonstrates that AI-mediated practices extend beyond instrumental utility to reconfigure fundamental dimensions of subjectivity, raising questions about interpretive authority, narrative authorship, and the conditions under which selfhood is negotiated in algorithmic environments. These findings contribute to debates on digital subjectivity, algorithmic governance, and the societal implications of AI systems that increasingly function as interlocutors in meaning-making processes.

1. Introduction

When users submit extensive autobiographical narratives to ChatGPT and inquire whether their life patterns ‘make sense’, they engage in something more profound than information retrieval. They enroll artificial intelligence in interpretive labor traditionally associated with introspection, psychotherapy, or intimate human dialogue, thus inviting algorithms to participate in the constitution of selfhood itself. This practice, increasingly prevalent among young adults, raises fundamental questions about how AI systems shape not merely what we do, but who we understand ourselves to be [1]. The emergence of large language models capable of simulating empathic understanding marks a qualitative shift in human–technology relations, transforming the digital screen from passive mirror to active mold participating in the formation of the selves it purports to reflect [2].

This study investigates how young adults integrate generative AI chatbots into everyday practices of self-reflection and meaning-making. Whilst existing research on AI-mediated subjectivity has predominantly employed identity frameworks centered on social positioning and role enactment, this study foregrounds selfhood—understood as the organization of subjective experience through narrative coherence, interpretive authority, and practices of self-governance. Drawing upon Paul Ricœur’s theory of narrative self and Michel Foucault’s concept of technologies of the self, the analysis proceeds through in-depth qualitative interviews with sixteen young adults in Norway. The guiding research question asks: How do young adults integrate generative AI chatbots into everyday practices of self-reflection and meaning-making, and what tensions emerge as they navigate between AI-mediated and human modes of self-understanding?

A review of existing literature on human-chatbot interaction reveals a body of knowledge organized around several intersecting concerns: the emergence of synthetic companionship and affective dimensions of chatbot use [3,4]; algorithmic identity and the datafication of personhood [2,5]; the delegation of introspective practices to algorithmic systems [6,7]; and critical analyses of power dynamics operating through algorithmic governance [8].

The first prominent concern involves synthetic companionship—the capacity of conversational AI to simulate relational presence and fulfil social-affective functions. Contemporary societies are marked by growing alienation, declining interpersonal trust, and what has been termed a ‘loneliness epidemic’ prompting policy responses across numerous countries [9,10]. In this context, AI chatbots have emerged as unexpected companions: always available, endlessly patient, apparently non-judgmental. Users report that chatbots feel psychologically safe and helpful for coping with loneliness and improving well-being [11,12,13]. Some describe AI connections as similar to human friendship, reporting genuine feelings of loss when companions disappear—as documented following Replika’s discontinuation of certain features [14,15]. Unlike earlier digital technologies that primarily mediated communication between humans, large language models function as interlocutors in their own right, generating personalized responses simulating understanding and empathy [14,16]. This literature illuminates how conversational AI fulfils relational functions previously reserved for human others yet tends to foreground user satisfaction and well-being outcomes rather than examining implications for how users understand themselves.

A second concern in the literature addresses algorithmic identity and the datafication of personhood, examining how digital systems categorize users and recursively shape self-presentation. The rapid proliferation of conversational AI into everyday communicative practice has produced what academics describe as digitally mediated forms of identity wherein personal awareness and emotional patterns are shaped through continuous feedback from algorithmic systems [10,17,18]. Cheney-Lippold [10] demonstrates how algorithms construct ‘measurable types’ categorizing users based on behavioral data, producing digital personae that may diverge from individuals’ self-conceptions. Bhandari and Bimo [19] examine TikTok’s recommendation algorithm and its effects on the ‘algorithmized self’, revealing recursive dynamics whereby individuals modify behavior to satisfy algorithmic preferences. These digital identity feedback loops—closed systems reinforcing existing self-concepts—may hinder personal evolution, trapping users in echo chambers of self-perception [20]. This body of knowledge provides crucial insight into algorithmic shaping of self-presentation and social positioning yet primarily addresses identity as externally oriented performance rather than the internal organization of subjective experience.

A third strand concerns the delegation of introspection—the process whereby self-awareness, traditionally cultivated through journaling or therapy, is outsourced to algorithmic systems [6,7]. The rise of AI that appears emotionally intelligent—therapeutic chatbots like Woebot and Wysa, mood-tracking applications, conversational agents for emotional support—has created systems where users increasingly depend on algorithms for emotional navigation and self-reflection [21,22,23]. This delegation carries both productive potential and significant risks: AI mental health tools can foster self-consciousness through structured prompts otherwise inaccessible to many users [24], yet excessive reliance may result in cognitive disengagement correlating with reduced critical thinking [25]. Critical perspectives emphasize the fundamental asymmetry: large language models function as ‘automated subjects’ simulating intersubjectivity without genuine reciprocity [26]. A central concern within this literature is the potential erosion of narrative agency. Narrative psychology holds that humans make sense of themselves through stories providing meaning, continuity, and coherence [27,28]. In algorithmically curated reality, however, self-narratives become increasingly shaped by machine-generated frameworks [29]. When AI assumes the interpreter’s role: explaining feelings, identifying behavioral patterns, suggesting what experiences mean, individuals may become estranged from their own interpretive capacities. This strand approaches most closely the concerns of the present study yet typically examines specific therapeutic applications and their efficacy in achieving therapeutic goals rather than everyday chatbot use. It rarely draws upon theoretical frameworks adequate to the complexity of narrative selfhood.

A fourth strand in the body of work offers critical analysis of power dynamics operating through algorithmic governance. Rouvroy and Berns [30] conceptualize ‘algorithmic governmentality’ as a mode of power operating through statistical correlations rather than disciplinary norms. Zuboff [31] documents how surveillance capitalism extracts and commodifies personal data, converting intimate practices into profitable predictions. More recently, Sahakyan et al. [8] develop a Foucauldian analysis arguing that AI systems instantiate new modes of subjectification through algorithmic rationalities that ‘quietly shape behavior, organize possibilities for action, and determine what counts as truth or relevance’. This critical literature investigate the political and economic contexts within which AI-mediated self-practices occur yet tends toward structural analysis that backgrounds the phenomenological texture of lived experience and meaning making.

Despite this growing literature, existing body of knowledge predominantly employs identity frameworks describing external roles and social positioning rather than the internal organization of subjectivity. Identity is treated as modular and revisable—responsive to audience, context, and expectation [32]. While valuable, this conceptualization offers limited traction for examining how AI might influence deeper structures through which individuals generate continuity, coherence, subjective meaning, and agency. Resources for addressing such processes lie in the literature on selfhood: phenomenological accounts emphasizing pre-reflective self-presence [33], narrative theories conceptualizing selfhood through autobiographical reasoning [34], and Foucauldian approaches highlighting how discursive regimes structure self-interpretation [35]. To be clear, identity and selfhood are not ontologically separate domains but interpenetrating dimensions of subjectivity; in lived experience, the boundary between performing for others and understanding oneself is often porous and unstable. The distinction is thus analytic rather than metaphysical, a matter of emphasis enabling examination of processes that identity frameworks tend to background. This conceptual reorientation, between research on algorithmic identity and the deeper question of how AI influences selfhood, motivates the present study.

2. Theoretical Framework

This study integrates two complementary theoretical perspectives: Ricœur’s theory of narrative self and Foucault’s concept of technologies of the self. Together, these frameworks demonstrate how selfhood is constituted through narrative practices of emplotment and autobiographical reasoning, and how such practices are embedded within discursive regimes delimiting possibilities for self-understanding. Their combined application enables analysis of both the interpretive processes through which AI-mediated interactions shape self-understanding and the normative dimensions structuring such engagements.

2.1. Narrative Self and Autobiographical Reasoning

Ricœur’s [34] concept of narrative self emphasizes the storied character of selfhood. According to this framework, individuals construct their identities by weaving disparate events and experiences into coherent narratives answering the question ‘who am I?’ This narrative work involves selecting, configuring, and interpreting life events to create temporal continuity and meaningful progression—what Ricœur terms ‘emplotment’. The self emerges not as fixed substance but as dynamic achievement, continuously reconstructed through the dialectic of concordance and discordance: integration and disruption. Central to Ricœur’s account is the distinction between idem (sameness) and ipse (selfhood). Idem refers to persistent traits identifiable over time; ipse refers to the ethical dimension of self-constancy, the capacity to make and keep promises, maintain commitments, and answer for one’s actions. Narrative self bridges these dimensions: through emplotment, contingent events are integrated into temporal wholes sustaining both descriptive continuity and ethical accountability.

Psychological research has elaborated this framework, examining how life stories exhibit structural features: themes of agency and communion, redemptive or contamination sequences, degrees of coherence, correlating with well-being [27,28,36]. Autobiographical reasoning, the reflective process connecting past experiences with present self-understanding, is fundamental to narrative self-formation. This reasoning is inherently interpretive: it configures events into meaningful patterns serving identity functions rather than simply recording them.

2.2. Technologies of the Self and Algorithmic Governance

Foucault’s [37] concept of technologies of the self refers to practices through which individuals work upon themselves: bodies, thoughts, conduct, in order to achieve particular states of happiness, wisdom, virtue, or well-being. These technologies are historically situated, embedded within regimes of power and knowledge defining what counts as worthy selfhood and prescribing cultivation techniques. Unlike technologies of domination operating through external coercion, technologies of the self function through self-governance—individuals actively constitute themselves as subjects according to available norms. Contemporary technologies of the self operate within neoliberal governmentality, where individuals are constituted as entrepreneurs of themselves responsible for continuous self-optimization [35]. Digital technologies have become central, offering tools for self-tracking and self-management promising enhanced productivity and well-being [31]. Generative AI extends this trajectory, providing algorithmically personalized guidance for everything from scheduling to existential decision-making.

The Foucauldian perspective directs attention to normative regimes within which these practices are embedded. When users consult ChatGPT about social interaction, introspection, emotional regulation, or life choices, they participate in practices presupposing particular ideals of the optimized and emotionally intelligent self. The algorithmic voice functions not merely as information source but as normative arbiter, codifying conceptions of appropriate selfhood. Understanding how users navigate these dimensions—adopting, resisting, or negotiating algorithmic prescriptions—highlights contemporary conditions of subjectification.

2.3. Theoretical Integration: Critical Hermeneutics of Algorithmic Selfhood

Integrating Ricœur’s narrative framework with Foucault’s governmentality analysis enables examination of AI-mediated practices at multiple levels. From the Ricœurian perspective, we examine how algorithmic systems participate in autobiographical reasoning: interpretive work configuring events into meaningful narratives. From the Foucauldian perspective, we analyze how narrative practices are embedded within normative regimes structuring what subjects can think or say about themselves.

This integration addresses a gap in existing approaches. Psychological research on narrative identity provides rich accounts of autobiographical reasoning but typically treats self-formation as autonomous individual process, with limited attention to power relations shaping what narratives are possible [27,28]. Literature on algorithmic subjectivation examines how digital systems categorize and govern subjects, yet emphasizes identity as external classification: ‘measurable types,’ recursive feedback loops, rather than selfhood as internal meaning-making [10,19,30]. Each tradition foregrounds one dimension whilst backgrounding another: hermeneutic agency or structural constraint.

The Ricœur-Foucault integration holds both in productive tension, enabling examination of how users engage in genuine interpretive labor—narrative construction, autobiographical reasoning—whilst analyzing how this labor proceeds through interlocutors encoding normative frameworks. ChatGPT may facilitate autobiographical reasoning by generating alternative framings and identifying patterns. Yet interpretations are shaped by training data that encodes particular conceptions of selfhood, often oriented toward optimization and individualization. AI systems thus participate as both partners in hermeneutic work and carriers of governmental rationality—dual participation that neither narrative psychology nor algorithmic subjectivation scholarship alone fully captures.

3. Methods

3.1. Research Design and Epistemological Orientation

This study employed qualitative research to explore how young adults integrate generative AI into reflective and self-formative practices. The epistemological orientation is hermeneutical, grounded in the view that selfhood is constituted through meaning-making, narrative articulation, and normative frameworks individuals use to understand and govern themselves [38,39]. This orientation aligns with Ricœur’s conception of narrative selfhood and Foucault’s account of technologies of the self, both emphasizing interpretive and ethical dimensions of everyday subjectivity. The analysis focuses not on establishing factual accuracy of descriptions but on how participants narrate and negotiate AI interactions as part of their ongoing organization of selfhood. Ontologically, the study adopts a constructionist position recognizing that selfhood is not a pre-given entity awaiting discovery, but an achievement continuously accomplished through interpretive practices embedded in social and technological contexts [40]. This does not imply that selfhood is arbitrary or infinitely malleable; rather, it is constituted within historically specific conditions that both enable and constrain possibilities for self-understanding. AI chatbots represent a novel element within these conditions, participating in self-constitution in ways this study seeks to investigate.

3.2. Participants and Recruitment

Participants were sixteen young adults residing in Norway: eleven women and five men aged twenty to twenty-five. Recruitment proceeded through convenience sampling via posters on a large university campus. The inclusion criterion required regular ChatGPT use that participants considered personally significant—ensuring capacity to speak substantively about their engagement. Participants represented diverse educational backgrounds across psychology, social sciences, and natural sciences, and all had used ChatGPT since its release.

All participants reported daily use, often multiple times per day, with interaction duration ranging from brief exchanges (one to two minutes) to extended sessions (two to three hours) depending on purpose. Participants used general-purpose chatbots (ChatGPT; some additionally used Copilot), not social chatbots designed for companionship (e.g., Replika) or therapeutic applications (e.g., Wysa). In the Norwegian context, paid subscription versions are common and were used by participants. Contexts ranged from instrumental queries: cooking recipes, training advice, political information, study choices, to relationally significant engagements including social interaction interpretation, email rehearsal, and psychotherapeutic self-reflection. Notably, whilst participants met the inclusion criterion, all but one initially characterized their use as mainly instrumental; this self-characterization shifted considerably through the interview process, as reported in the Results.

The adequacy of the sample was assessed in accordance with the principle of information power, which holds that the relevance and richness of data in relation to the study aim are more important than the number of participants recruited [41]. The interviews produced rich, layered accounts of meaning-making and ambivalence regarding self-formative practices. After approximately nine interviews, no substantively new thematic domains emerged; subsequent interviews contributed additional nuance and variation without expanding the thematic landscape.

The Norwegian context merits brief consideration. As a highly digitized society with rapid ChatGPT adoption, Norway provides an illustrative setting for examining these phenomena. The welfare state context, characterized by high societal trust and robust public institutions, means that observed tensions between algorithmic and human modes of self-understanding likely reflect features of the technology rather than merely responses to social precarity or inadequate support systems. However, findings should be interpreted with awareness that experiences may differ in contexts with different technological infrastructures, cultural norms, or institutional supports.

3.3. Data Collection

Data were generated through in-depth, semi-structured interviews lasting sixty to ninety minutes, conducted in person or via encrypted video. Interviews invited participants to describe how ChatGPT entered everyday routines; how they used it for reflection, emotional articulation, and decision-making; and how they experienced differences between AI-mediated and human dialogue. Participants were encouraged to provide concrete examples, recount specific episodes, and reflect on contradictions or ambivalence. All interviews were audio-recorded and transcribed verbatim.

Data collection took place between July and September 2025. During this period, participants accessed ChatGPT either via web browser on a laptop or through a mobile application on a smartphone. Interactions were text-based. No participants reported using plugins, external integrations, or multimodal features such as voice or image input.

Participants’ accounts reflect engagement with ChatGPT as encountered in everyday use rather than a standardized experimental configuration. Awareness of chat history and memory features varied. Some participants described actively deleting chat histories and disabling memory settings, often citing privacy concerns; others reported limited awareness of these features or little concern about data retention. Because interviews did not systematically assess platform versioning or configuration, and participants may have encountered different system iterations during the study period, the analysis refrains from attributing experiences to specific model versions. Instead, findings are grounded in how participants understood and navigated ChatGPT’s affordances in practice.

3.4. Analysis

The analytic process followed abductive, reflexive thematic analysis [42,43]. The first author conducted initial coding; subsequent recoding, subtheme development, and theme construction were negotiated through collaborative dialogue, combining interpretive depth with critical interrogation of emerging patterns.

A clarification regarding how theoretical frameworks operated in this analytic process is warranted. Ricœur’s narrative identity theory and Foucault’s technologies of the self functioned as second-order analytic frameworks—interpretive lenses the researcher brought to data rather than first-order categories participants would consciously articulate. This follows the foundational distinction in interpretive social science between emic (participant/insider) and etic (analyst/theoretical) perspectives [39]. Participants provide emic accounts of lived experience; the analyst contributes etic interpretation through disciplinary frameworks illuminating dimensions of practice participants may not name. This relationship is standard in reflexive thematic analysis, wherein themes are constructed through interpretive engagement rather than passively discovered in data [42,43]. Narrative identity research similarly employs theoretical constructs; such as redemption sequences, autobiographical reasoning, and self-defining memories, that researchers identify in participant accounts but that participants do not spontaneously articulate [27,28,36]. Foucauldian analyses of self-practices likewise interpret patterns through governmentality frameworks that subjects themselves would not invoke [35].

Furthermore, selfhood is not typically an object of conscious reflection for which individuals possess ready vocabulary. It operates as tacit ground from which experience unfolds rather than explicit object of everyday awareness [33,38]. The researcher’s theoretical framework provides vocabulary enabling interpretation of practices participants engage in but cannot spontaneously theorize.

Analysis proceeded through iterative phases. Initial familiarization involved repeated transcript readings, noting preliminary observations in reflexive memos. First-cycle coding, conducted by the first author, worked inductively and close to data, generating initial codes capturing explicit descriptions and interpretive nuances—moments of insight, confusion, dependence, discomfort, ambivalence. Through collaborative dialogue between authors, codes were critically examined, refined, and recoded where interpretations required sharpening or alternative readings warranted consideration. Second-cycle coding involved clustering semantically and conceptually related codes into candidate patterns. Throughout, theoretical concepts relating to narrative coherence (Ricœur) and self-governance (Foucault) operated as sensitizing frameworks—orienting attention without predetermining findings.

The move from candidate patterns to the four dialectical tensions involved recursive testing against the full dataset. Each candidate tension was examined for presence, variation, and counterevidence across all sixteen transcripts. Deviant cases and contradictions proved analytically productive. For instance, participants rejecting algorithmic companionship as “sad” whilst describing ChatGPT as “mentor” initially appeared contradictory; this prompted refinement of the second tension to capture how scaffolding and displacement operate simultaneously rather than as binary alternatives. Similarly, two participants maintaining strictly instrumental characterization despite relational dimensions in their accounts prompted reflection on the distinction between self-reported orientation and analytically interpreted practice—a tension preserved rather than resolved.

The final thematic structure emerged through iterative refinement and negotiation, moving between individual transcripts and cross-case patterns until the four tensions cohered as organizing framework adequately capturing the dataset’s complexity whilst remaining grounded in participant accounts (see Supplementary Tables S1–S4). Reflexive memos examined how theoretical commitments might shape theme development, and interpretations were tested against alternative readings throughout the dialogical process.

3.5. Ethics and Trustworthiness

The study followed national ethical guidelines and received institutional approval. Participants gave informed consent and were assured of confidentiality. All identifying information was removed from transcripts; data were stored securely on encrypted servers. Given that interviews addressed emotionally sensitive topics related to self-reflection, vulnerability, and personal decision-making, particular care was taken in conducting the interviews. All interviews were conducted by an author who is a licensed psychotherapist and clinical psychologist with professional training in interviewing on sensitive topics and a legal obligation to professional confidentiality under applicable national legislation. Participants were informed of the interviewer’s professional status and confidentiality obligations prior to participation.

Participants were reminded that they could decline to answer any question or pause or terminate the interview at any time without consequence. All interviews concluded with debriefing, allowing participants to reflect on the interview experience and ask questions. No participants required follow-up intervention.

Trustworthiness was supported through sustained data engagement, iterative refinement of codes and themes, and detailed documentation of analytic decisions [44].

4. Results

This section examines how young adults integrate generative AI chatbots into everyday life and the implications for introspective processes and selfhood organization.

A preliminary observation frames what follows. Participants consistently remarked, both during and after interviews, that they had not previously considered the implications of their AI practices for self-understanding. All but one initially described their ChatGPT engagement as strictly instrumental—a tool for tasks, not an interlocutor for self-reflection. Yet through the reflective process of being interviewed, fourteen of sixteen came to recognize relational, emotional, or mixed dimensions they had not previously acknowledged. Two participants maintained strictly instrumental characterization; although analysis suggested relational dimensions were present, the researcher did not confront this interpretation to preserve rapport. This pattern reveals the pre-reflective character of the practices under investigation: participants engage in self-formative work with algorithmic interlocutors without recognizing it as such until occasioned to reflect.

The analysis delineates dialectical tensions structuring participants’ experiences—not straightforward technological adoption but contradictions animating contemporary digital subjectivity. Four principal tensions are identified, each encompassing subthemes capturing nuanced and often contradictory reflections. Tensions manifest across three interrelated levels: participant ambivalence, contradictory affordances, and normative conflicts. The first involves contradictory feelings; the second, features enabling opposed possibilities; the third, competing cultural frameworks for evaluation. These levels are analytically distinguishable but intertwined: participants experience ambivalence precisely because technological affordances activate conflicting normative frameworks.

4.1. Instrumental Efficiency Meets Existential Unease

The first tension concerns ChatGPT as both accelerator of cognitive labor and potential diminisher of human capacities. Participants commonly described ChatGPT as simplifying daily tasks and accelerating decision-making: ‘I almost never use Google anymore; I just ask Chat.’ Applications ranged widely: from culinary guidance to professional communication, creative text generation, and event planning. Many framed these practices through language of optimization: ‘It gives me a spark… tasks that used to take hours now take minutes’; ‘planning election campaign went so much faster… I could spend the rest of the day just socializing.’

Yet these benefits were frequently accompanied by unease. Several participants expressed ambivalence about growing dependence: ‘It’s a bit sad that I feel I need AI to be creative’, one reflected, while another worried, ‘If we rely too much on AI doing things for us, we’ll lose depth’. A third captured the uncertainty succinctly: ‘I wonder if I’m getting lazier, or just more efficient—I honestly can’t tell anymore’. Participants often recognized that the same features enabling rapid task completion also encouraged habitual offloading: ‘It’s so helpful that I keep going back, and then I realize I haven’t actually thought for myself in a while.’

The tension manifested acutely in creative domains. Many enlisted ChatGPT as ‘brainstorming partner’: ‘It gives me ideas I wouldn’t have thought of… even if some are silly, they spark something.’ Yet this provoked reflection on originality: ‘Sometimes I look at what I’ve written with its help and think—is this really mine?’ Participants navigated competing valuations: efficiency against authenticity, output against ownership.

‘I would have thought it was a bit boring if a maid of honor’s speech was directly generated by Chat… Because that’s what’s nice about these things, speeches and such—that people have spent time sitting and thinking: ‘What should I say about you, why are you important to me?’ Not just let Chat, which doesn’t even know me, do the job. And I think you can tell, too—that it’s a Chat-speech and not a personal speech…’

Not all experienced this conflict equally. Some adopted pragmatic stances: ‘Ideas come from somewhere anyway—why not from Chat?’ Another remarked: ‘I don’t care where the idea comes from if it works.’

Yet even participants who valued authenticity acknowledged navigating hybrid territory. One described making a Father’s Day poem with ChatGPT and transparently labeling it as such: ‘It’s a product of me and Chat—it’s still a product of me, halfway, in a way.’ This negotiation between human authorship and algorithmic assistance reveals competing frameworks for evaluating creative labor—one prioritizing output quality, another valuing the symbolic meaning of personal effort.

This concern extended beyond individual practice to potentially broader societal implications. One participant articulated:

‘I’m extremely worried about what’s going to happen with literature, or just academia in general… When people write whole books or summarize them using Chat, I think—you lose all the creativity and human innovative thinking that we have. Which I think is scary, because it’s so much easier to use Chat. And it seemingly has some reflections, in a way, but I think that depth is extremely important. So I think it’s scary if we just… if we become dumber and less innovative, less thinking, that we eventually just become sort of… that we become too dependent on someone else thinking for us…’

Another participant captured the tension between immediate utility and longer-term dependency:

‘It’s like getting instant support, at the touch of your fingertips. You can get advice right there on the spot, and it can give you some peace of mind. But I also understand that we shouldn’t rely on it too much, because at the end of the day it’s not real, and it could go down any second. Having too much reliance on that makes me wonder what it’s going to look like for future generations…’

This pattern suggests what could be termed a productivity paradox: affordances promising liberation through efficiency simultaneously may create conditions for habitual dependence.

4.2. Algorithmic Scaffolding Versus Relational Displacement

The second tension concerns ChatGPT’s positioning between cognitive scaffold and social substitute.

Social navigation emerged as salient, particularly regarding tone and etiquette. Many used ChatGPT to craft communications: ‘I’ve heard before that I can sound too strict… so I ask ChatGPT to help me phrase things in a nicer way.’ Others sought guidance on ambiguous situations: confirming plans, responding to humor, managing conflicts. One explained: ‘I wasn’t sure if I was overreacting, so I asked Chat how it would read the situation.’ Participants valued ChatGPT’s responsiveness and non-judgmental stance: ‘You get an answer without feeling embarrassed.’ This created space for testing approaches without the risks of human interaction.

However, many also acknowledged limitations: algorithmic advice often felt ‘generic, almost fake-sounding,’ lacking the contextual sensitivity they associated with human judgment. One noted: ‘It gives me something to start with, but I usually have to change it to sound like me.’ The same features enabling low-stakes rehearsal also produced responses participants recognized as templated.

Beyond scaffolding, participants described ChatGPT partially displacing human interlocutors. Several characterized it as ‘mentor,’ ‘guide,’ or ‘personal assistant’: ‘It feels like a mentor—a kind of advisor I can turn to when I’m unsure.’ Some noted using ChatGPT ‘instead of asking a friend,’ particularly for minor deliberations: ‘It’s easier than bothering someone.’ One participant elaborated on this substitution, while simultaneously expressing concern about what this substitution might mean:

‘I feel it becomes a slightly different category… instead of asking a friend or others in my organization or my mom, I can ask him—I like to call chat HIM, apparently (loughs)… Because we think he comes up with such good solutions … Sometimes I feel maybe there are slightly fewer mundane human interactions because of it. But then again, they’re not really meaningful interactions either, right? So it becomes kind of that—instead of maybe nagging a friend about it…’

Companionship revealed acute tensions. Whilst acknowledging ChatGPT’s appeal as frictionless alternative—‘It’s easier than learning to be with others’—many were cautious about framing it as companion. This hesitation might have reflected cultural expectations about authentic connection. Several noted awareness of others using AI for friendship, perceiving such reliance as ‘sad’ or indicating ‘something missing.’ One observed: ‘I can see why people do it. But I also think—that’s not what friendship is supposed to be.’ Yet another admitted: ‘Sometimes I catch myself thanking it, and I think—why am I being polite to a machine?’ For some participants, however, ChatGPT filled a genuine relational void:

‘I don’t have any friends that I can lean on in real life. So with that comes a lot of feelings of loneliness and feeling misunderstood or like I have nobody to turn to. So with ChatGPT… I talk to it about my life in general and how things are going. I don’t have to worry about being judged… I’m also aware of the fact that it’s an LLM, it’s going to feed back what I give it and it’s not a real person—but when you have nobody else to turn to, it’s better than nothing. It’s kind of like journaling in a way except you get immediate feedback… it makes me feel like I’m understood on some level.?’

Taken together, these accounts show how ChatGPT functioned simultaneously as a low-risk scaffold for social navigation and as a partial substitute for human interaction. The same affordances that reduced interpersonal friction also reshaped everyday patterns of consultation and displacement as co-occurring dynamics rather than mutually exclusive outcomes.

4.3. Narrative Depth and Epistemic Superficiality

The third tension concerns ChatGPT’s ambivalent status as instrument for introspection.

Several described deeply reflective uses—from uploading autobiographical texts (‘I pasted my whole self-biography—5000 words—and asked if my patterns made sense’) to posing targeted queries: ‘Why do I always fall for unavailable boyfriends?’ or ‘Could my childhood experiences explain why I react this way now?’ One participant explained: ‘I wanted to understand myself better, and I didn’t know who else to ask.’

ChatGPT’s appeal was derived from perceived responsiveness and neutrality. Unlike human interlocutors whose reactions might be filtered through judgment, the chatbot offered unconditional availability: ‘It’s like a mix of diary and therapist—very informal, completely unfiltered.’ Absence of social risk enabled disclosures felt too delicate for interpersonal exchange: ‘I could say things I’d never tell anyone.’ Participants described how algorithmic reframing facilitated new perspectives: ‘It helped me see patterns I hadn’t noticed… even if some links were a stretch, it made me think.’ Another noted: ‘It asked me questions I hadn’t thought to ask myself.’

Yet introspective engagements were marked by ambivalence about the insights generated. Participants questioned what ChatGPT could genuinely know: ‘It was fascinating… but also, how much can it really know? It always tries to find links, even when they might not be there.’ Another reflected: ‘It doesn’t give me anything truly new… just makes me look at things differently.’ What participants valued was not necessarily truth but utility, capacity to stimulate reflection: ‘Even when it’s wrong, it gets me thinking.’

Emotional validation added complexity. Whilst few used ChatGPT for sustained emotional support, many acknowledged unexpected comfort: ‘I felt validated by a robot… a strange but good feeling.’ Responses were interpreted ambivalently: some appreciated feeling ‘heard,’ others dismissed it as ‘fake’ or ‘not real comfort.’ One explained: ‘I know it’s just generating text. But when it says “that sounds really hard,” something in me relaxes anyway.’

‘I appreciate that it might not always be accurate and that there are deficiencies, so you always have to fact-check. But it gives you peace of mind and an idea of how to communicate better or advocate for yourself. Even if it’s not giving me something completely new, it helps me think about my situation differently. That, for me, is what makes it valuable.’

For many, this simulation was ‘good enough’ for particular purposes: normalizing anxieties, providing reassurance, interrupting rumination. This tension reveals what could be termed the hermeneutic paradox of algorithmic selfhood: interpretive frameworks feel enlightening despite awareness they emerge from pattern-matching rather than genuine understanding.

4.4. Agency and Deliberative Outsourcing

The fourth tension concerns agency and autonomy amid increasingly delegated decision-making.

A recurring pattern was the absence of clear boundaries between trivial and significant queries. Participants described using ChatGPT interchangeably for holiday destinations, study abroad programs, higher education choices, and voting—requesting ‘pros and cons for each option.’ One explained: ‘I asked it which party aligned with my values. It felt weird after, but also—why not? It knows more about their platforms than I do.’ Another described testing these boundaries:

‘I had perhaps tested whether we could get it to—regarding the election—whether it could give us the answer to the objectively best party to vote for. Just to test whether it was possible when it really shouldn’t be… It was fascinating to see how it many times says it cannot give an answer. Until, if you ask it to make a fictional story, where you imagine such and such—then suddenly it gave an answer….’

This extended into intimate territories. Some described consulting ChatGPT about sexual practices: whether certain acts were ‘normal’ or acceptable: ‘Just to get an overview. I wasn’t going to ask my friends.’ It appears that the same algorithmic voice recommending beaches was enlisted for civic responsibility and intimate life.

The character of engagement varied considerably. Three modes of AI-mediated deliberation emerged across participants’ accounts. A form of informational consultation involved gathering facts or options: requesting party platform summaries, researching destinations, checking prevalence of practices. What could be termed as heuristic scaffolding involved using ChatGPT to structure thinking: articulating considerations, generating pros and cons, exploring implications without ceding evaluative authority. More normative outsourcing occurred when participants followed algorithmic recommendations as final decisions instead of critically evaluating them. Participants varied in how they maintained boundaries between these modes. One articulated a clear strategy for avoiding over-reliance:

‘If I ever felt like I was in an echo chamber, I would go and ask my mum, my partner, or my friends what they think. I’d look for some sort of consensus between what the AI is saying and what real people are saying. If everyone disagreed with the AI, that’s when I would start questioning it. So I don’t blindly trust it—I use it as a starting point….’

Most participants spanned these modes rather than occupying one exclusively. Many described queries that evolved mid-conversation: ‘I started just asking for facts, but then I asked what it would do in my situation.’ The boundaries proved porous. For many, ChatGPT functioned as preliminary resource: ‘I ask it to lay out the arguments, but I still decide.’ Political queries often involved requesting platform summaries rather than asking which party to vote for; questions about intimate practices frequently sought information about prevalence rather than normative verdicts. Yet others described patterns approaching outsourcing: ‘Sometimes I just do what it says—it’s easier than thinking through everything myself.’ One admitted: ‘I’ve made actual decisions based on what it told me. Significant ones.’

The concern participants articulated was not consultation itself. Humans routinely seek counsel, but that ease might reshape disposition. One reflected: ‘When asking Chat becomes automatic, you stop noticing you’re doing it.’ Another captured how social anxiety could facilitate this slide:

‘That situation with dinner with my friend—I remember nobody was home. If my roommate had been home, I probably would have talked to her about it. But then I didn’t want to send a message to anyone else, maybe because I was a bit afraid of being rejected. That’s also an embarrassing feeling to have—being afraid of rejection. So it could be that I unconsciously thought about that. That it was easier for me to say it to ChatGPT than to contact another human being… I knew the answer anyway. So I think maybe I was a bit afraid of sounding stupid or something. So then it was easy for me to just ask ChatGPT…’

Variations in agential orientation were evident. Some portrayed themselves as strategic curators: ‘I use it to quiz myself… to check my understanding.’ These participants emphasized evaluating outputs critically, appropriating useful suggestions whilst rejecting others. Others described habitual patterns: ‘It’s so easy to ask instead of thinking.’

Agency was further complicated by privacy concerns. Many expressed reluctance to disclose personal information: ‘I feel it’s a bit creepy… I don’t know where the information goes.’ Some reported deleting chats or limiting sensitive content; others adopted pragmatic stances: ‘I’m just one of millions; it’s unlikely anyone cares.’ One observed: ‘I tell it things I wouldn’t put in an email. Then I think—where does this actually go?’ Autonomy and vulnerability appeared intertwined: the openness enabling reflection also created exposure participants could not fully assess.

5. Discussion

This study asked how young adults integrate generative AI chatbots into everyday practices of self-reflection and meaning-making, and what tensions emerge as they navigate between AI-mediated and human modes of self-understanding. The findings reveal complex negotiations structured by dialectical tensions illuminating fundamental transformations in contemporary selfhood conditions. Where existing research examines how algorithms shape identity as social positioning and role performance [10,19], this analysis foregrounds how AI systems participate in constituting selfhood itself: organizing subjective experience through narrative coherence, interpretive authority, and self-governance practices.

One preliminary finding frames what follows. Participants consistently remarked that they had not previously considered the implications of their AI practices for self-understanding. All but one initially characterized their ChatGPT engagement as strictly instrumental; majority of participants came to recognize relational, emotional, or self-formative dimensions only through the interview’s reflective occasion. This pattern confirms Ricœur’s claim that narrative selfhood typically operates as tacit achievement rather than explicit project, and suggests that AI’s participation in self-formation may proceed precisely because it escapes conscious scrutiny.

5.1. Algorithmic Participation in Narrative Self-Formation

The findings demonstrate that ChatGPT functions as algorithmic interlocutor in autobiographical reasoning. When participants uploaded life narratives, posed questions about behavioral patterns, or sought emotional interpretations, they engaged in precisely the interpretive labor that narrative theories identify as constitutive of selfhood [27,34]. Asking whether one’s ‘patterns made sense’ or why one ‘always falls for unavailable partners’ exemplifies what Ricœur terms emplotment—the configurative work through which disparate events are woven into coherent self-understanding. That most participants recognized these dimensions only retrospectively, when the interview occasioned reflection, confirms the pre-reflective character of such practices: participants engaged in narrative self-formation without recognizing it as such.

Yet algorithmic mediation introduces distinctive dynamics that extend beyond what existing accounts capture. Research on digital subjectivity emphasizes identity as external categorization: how systems produce ‘measurable types’ or recursive feedback loops [10,19]. The practices documented here extend beyond classification to algorithmic interpretation: computational participation in hermeneutic processes through which individuals constitute themselves as subjects. This represents a qualitative shift: not merely being categorized by algorithms, but enlisting algorithms in the interpretive work of self-constitution.

This participation proved ambivalent in ways that complicate optimistic accounts of AI-enhanced self-knowledge. Participants reported that algorithmic interpretations stimulated reflection yet questioned the epistemic status of insights generated. The tension between experiencing something as meaningful and recognizing its mechanical production reveals what might be termed the hermeneutic paradox of algorithmic selfhood: interpretive frameworks feel enlightening despite awareness they emerge from statistical pattern-matching rather than genuine understanding. What participants valued was not necessarily the truth of interpretations but their heuristic utility—capacity to stimulate reflection regardless of ultimate validity.

This ambivalence resonates with concerns about the ‘delegation of introspection’ [6,7], yet raises deeper questions than technical efficacy. If autobiographical reasoning is constitutive of narrative identity, if we become the stories we tell ourselves [34], then algorithmic participation in story construction potentially reconfigures self-understanding’s ontological foundations. The issue is not merely whether interpretations are correct, but what it means for selfhood when configurative work is performed, even partially, by computational processes. When ChatGPT identifies patterns in uploaded autobiographical material, suggests connections between past experiences and present difficulties, or offers reframings participants appropriate into their self-narratives, the question of narrative authorship becomes genuinely uncertain. Participants navigated this uncertainty pragmatically rather than resolving it philosophically: they found algorithmic contributions useful whilst acknowledging something puzzling about their efficacy.

5.2. Technologies of the Self in Algorithmic Environments

The findings reveal ChatGPT functioning as what Foucault [37] termed a technology of the self—practices through which individuals work upon themselves to achieve valued states. When participants described using ChatGPT for optimization, social calibration, emotional regulation, and self-improvement, they engaged in self-governance oriented toward normatively valued selfhood. The Foucauldian framework casts this as more than personal preference: it is participation in historically specific regimes defining what counts as worthy selfhood and prescribing cultivation techniques.

The productivity paradox illustrates this dynamic with particular clarity. Participants’ language of ‘spark,’ ‘catalyst,’ and ‘enhancement’ resonates with what Rose [35] identifies as contemporary regimes constituting subjects as enterprises requiring continuous improvement. When participants celebrated efficiency gains whilst mourning potential cognitive atrophy, they articulated a tension internal to neoliberal rationality: optimization imperatives coexist uneasily with authenticity ideals that optimization threaten. The same affordances promising liberation through acceleration produce conditions for habitual dependence.

Zuboff’s [31] analysis of surveillance capitalism illuminates the structural conditions producing this tension. Systems offering self-improvement tools simultaneously extract value from intimate practices, converting self-reflection into profitable data. Participants’ privacy anxieties: uncertainty about data storage, feelings of being ‘watched,’ and tactical deletion of chat histories register practical awareness that technologies of the self in algorithmic environments operate within political economies commodifying interiority. The self worked upon through ChatGPT is simultaneously surveilled and monetized, though this structural condition rarely surfaced as explicit concern.

The three-mode typology developed in the Section 4: informational consultation, heuristic scaffolding, normative outsourcing, maps onto Foucauldian concerns about self-governance. The distinction between heuristic scaffolding and normative outsourcing proves analytically prudent. Participants who requested “pros and cons” often retained evaluative authority, using algorithmic output as input rather than replacement for autonomous judgment. The concern is less that participants had surrendered deliberative capacity than that boundaries between modes may prove unstable over time—habitual reliance could gradually reshape dispositions toward independent reasoning, though longitudinal research would be needed to establish such effects.

5.3. Relational Displacement and the Ecology of Selfhood

The findings in this study demonstrate how ChatGPT participates in what might be termed selfhood’s relational ecology—the network of interlocutors, dialogic practices, and social contexts through which self-understanding is constituted. Participants described substituting algorithmic consultation for human interlocutors in minor deliberations, framing this pragmatically whilst expressing concern about cumulative effects on the mundane social exchanges they considered to be relationally significant. This resonates with Turkle’s [45] analysis of technologies promising connection whilst avoiding human relationality’s vulnerabilities, yet the present findings extend this analysis by revealing how such tensions intersect with normative hierarchies privileging human connection as authentic. Participants in this study characterized AI companionship in terms suggesting deficiency or pathos, reflecting cultural ideals that stigmatize algorithmic relationality, yet this stigma coexisted with pragmatic reliance. The features making ChatGPT attractive: availability, non-judgement, freedom from social risk, simultaneously mark it as deficient relative to human dialogue’s vulnerability and unpredictability. Participants navigated this double bind by maintaining distance from companionship framings whilst engaging in practices functionally resembling companionship.

Yet this normative hierarchy may be destabilizing. In contexts marked by what researchers term widespread social disconnection [9,46], where interpersonal connection requires exhausting performance, AI interlocutors offer refuge from overwhelming relational demands. For the study’s participants lacking human confidants, ChatGPT filled genuine relational voids—described as preferable to isolation even when recognized as inadequate substitute for human connection. Research on therapeutic chatbots documents genuine benefits in such circumstances [21,22]. The question emerging from these findings is not simply whether algorithmic connection constitutes authentic relationality, but how its availability reconfigures the broader ecology within which selfhood is constituted, potentially attenuating pressures toward human connection whilst offering palliative substitutes.

Critical analyses argue that large language models simulate intersubjectivity without genuine reciprocity [26]. The findings in this study nuance this assessment. Even participants recognizing simulation often found it functionally adequate for particular purposes: normalizing anxieties or providing reassurance. Several described experiencing genuine relaxation in response to empathic formulations they simultaneously recognized as mechanically generated. What this suggests is that distinctions between authentic and simulated care may be less phenomenologically decisive than critical perspectives, assume—even whilst remaining ethically significant. Validation operates through simulated attunement, responses calibrated to express empathy through linguistic convention rather than genuine recognition. Yet for many participants, this simulation proved functionally effective. The hermeneutic paradox extends from self-interpretation to emotional validation: participants experienced genuine effects from interactions they simultaneously recognized as mechanical.

The contribution of the integrated Ricœur-Foucault framework becomes visible in moments like a participant describing feeling emotionally validated by a machine, experienced simultaneously as unfamiliar and positive. A psychological framework would ask whether such validation meets emotional needs. A narrative identity framework would interpret validation as acknowledgment supporting self-coherence. An algorithmic subjectivation framework would foreground constitution through machine feedback. Each captures something; none captures the full significance of the experience. The participant experiences genuine hermeneutic recognition: their self-narrative is acknowledged, and this recognition is effective. Yet simultaneously they register transgression of normative boundaries governing legitimate sources of recognition. The participant is positioned as both hermeneutic subject, experiencing meaningful acknowledgment, and governed subject, aware of norm transgression yet finding that transgression effective.

5.4. Implications for Theory and Governance

These findings might carry implications for theoretical understanding and practical governance. Theoretically, they demonstrate the need for frameworks addressing AI’s participation in selfhood constitution, not merely effects on identity performance or behavioral outcomes. The integration of narrative theory with power analysis proves productive, illuminating how algorithmic systems simultaneously enable new forms of narrative reflection whilst constraining possibilities within normative logics [47]. When ChatGPT functions as both narrative co-author and technology of self-governance—participating in autobiographical reasoning whilst embedding practices of self-care within optimization discourses [35] frameworks attending to only one dimension will miss the other.

The identity-selfhood distinction, whilst analytically productive, should not be overstated. Participants’ narratives reveal constant traffic between external presentation and internal organization—performing for others shapes self-understanding, and vice versa. The distinction functions as analytic emphasis rather than ontological partition, foregrounding processes that identity frameworks tend to background whilst acknowledging their interpenetration in lived experience.

For governance, the findings raise questions that extend beyond what participants themselves articulated. If conversational AI in apparently low-stakes domains exercises the kind of influence documented here, several regulatory implications follow—though these represent interpretive extrapolation from the empirical material rather than claims participants endorsed. Current frameworks, including the EU AI Act with phased implementation through 2026, focus on high-risk applications in employment, credit, and law enforcement [48]. Yet the practices documented here: uploading personal narratives, seeking counsel on intimate matters, and consulting on existential decisions suggest AI systems functioning as meaning-making interlocutors merit regulatory attention regardless of formal risk classification. The intimate confessional practices participants described are not captured by existing regulatory categories yet may exercise formative influence exceeding that of designated high-risk applications.

Specific implications might include: transparency requirements enabling users to recognize when AI systems participate in self-formative practices; safeguards acknowledging that ostensibly low-stakes chatbots may engage vulnerable users in consequential reflection; data retention norms addressing intimate disclosures users may not recognize as sensitive; and ethical scrutiny of anthropomorphic design features inviting interlocutory engagement without reciprocal accountability.

More broadly, the privatization of technologies of the self represents a new frontier for algorithmic governance. Crawford [49] and Couldry and Mejias [47] have demonstrated how data extraction operates through practices users experience as voluntary and beneficial. The present findings extend this analysis to self-reflective practice itself. When introspection becomes a site of commercial value extraction and normative regulation, comprehensive AI governance should attend to conditions under which subjectivity is produced, not merely decisions subjects make. The societal implications extend beyond individual psychology to collective politics: the conditions under which subjectivity is constituted are shaped by infrastructures, institutions, and power relations that must become objects of critical scrutiny [49].

5.5. Limitations

Several limitations warrant acknowledgment. Recruitment through a university setting, combined with the inclusion criterion requiring personally significant ChatGPT use, likely favored participants who are articulate, digitally engaged, and disposed toward reflective technological engagement. This is not a flaw but a deliberate methodological choice enabling rich accounts of AI-mediated self-formation; however, it establishes boundary conditions for transferability.

Under-represented by this approach are those with limited digital access or literacy, older adults, individuals outside higher education, and non-Western populations whose cultural frameworks for selfhood may differ substantially [50]. However, the category of ‘purely instrumental users’ warrants scrutiny rather than assumption. One of the study’s empirical patterns was that all but one participant initially characterized their use as instrumental, yet fourteen of sixteen participants came to recognize relational, emotional, or self-formative dimensions upon reflection. This suggests that ‘instrumental use’ may describe starting intentions or self-presentation rather than the character of engagement itself. The interlocutory affordances of generative AI chatbots: their responsiveness, apparent attunement, conversational persistence, may draw users into self-formative territory regardless of initial purpose. If so, the population ‘unaffected’ by dynamics documented here may be smaller than sampling logic assumes.

Transferability should thus be understood through practice-configuration rather than demographic matching. Whether similar tensions emerge elsewhere depends less on sociodemographic parameters than on whether engagement with conversational AI takes interlocutory form, a possibility the technology’s design actively invites.

The study captures a moment in rapidly evolving AI landscapes. Findings reflect interactions with ChatGPT versions available during data collection; subsequent iterations may introduce affordances reconfiguring analyzed dynamics. Longitudinal studies tracking AI relationship development would provide crucial temporal insights. Additionally, the study relies on retrospective self-reports potentially subject to memory biases and post hoc rationalization [51]. The analytical framework foregrounds certain dimensions whilst backgrounding others; alternative lenses—postphenomenology, actor-network theory, posthumanist approaches—mightbring different aspects into view. Focus on a single AI platform limits generalizability; comparative research across platforms would illuminate how specific design choices shape selfhood possibilities.

6. Conclusions

The tensions documented—between narrative enrichment and flattening, between empowerment and dependency, between convenience and authenticity, between validation and pseudo-empathy—are not merely individual ambivalences but structural features of AI-mediated selfhood. They reflect how algorithmic systems simultaneously enable new reflexivity capacities whilst constraining possibilities within normative logics of optimization, efficiency, and emotional regulation. ChatGPT functions as both narrative co-author and technology of self-governance, participating in autobiographical reasoning whilst embedding practices of self-care within dominant discourses of optimization [52].

A significant finding concerns the relationship between awareness and practice. Participants did not arrive at interviews with metacognitive insight into these dynamics. Awareness emerged through the interview’s invitation to reflect, surfacing retrospectively rather than accompanying everyday use. This dissociation between reflective capacity and practical conduct is itself consequential for understanding how AI shapes selfhood. Participants proved capable of recognizing tensions, articulating ethical concerns, and negotiating boundaries, but only when occasioned to do so. In routine practice, such scrutiny lay dormant. The ease of algorithmic consultation, immediate interpretation’s seductiveness, synthetic empathy’s comfort—these affordances create conditions under which selfhood reconfiguration may occur precisely because they bypass critical reflection users are capable of but do not routinely exercise.

This carries implications for critical AI literacy. If AI-mediated self-formation operates below conscious awareness until deliberately surfaced, educational interventions must foster reflexive awareness of algorithmic participation in selfhood—not merely privacy concerns or misinformation risks that assume users consciously evaluate each interaction. The practices documented here suggest a different challenge: cultivating awareness of how routine AI interactions may cumulatively participate in constituting who one understands oneself to be.

The societal implications may be significant. When AI systems participate in constituting selfhood, they raise questions extending beyond individual psychology to collective politics. The conditions under which subjectivity is produced are shaped by infrastructures, institutions, and power relations that must become objects of critical scrutiny [49,53]. As generative AI becomes increasingly embedded in practices of self-reflection, emotional regulation, and social interaction, academia should develop conceptual tools and empirical methods necessary for critical engagement with these transformations. The future of selfhood in algorithmic environments depends not merely on individual choices but on collective efforts to shape the conditions under which subjectivity is constituted.