Unmasking AI: Discovering The Unexpected Connection To Our Identity

Unmasking AI: Identity, Self-Perception, and the Emergence of AI-Induced Impostorism

Simone S. Atkinson
The Atkinson Institute · theatkinsoninstitute.org
Peer-Informed Article · 2025
Topics: Impostorism · AI & Identity · Self-Worth · Social Comparison

Suggested Citation (APA):
Atkinson, S. S. (2025). Unmasking AI: Identity, self-perception, and the emergence of AI-induced impostorism.
The Atkinson Institute. https://theatkinsoninstitute.org/research/unmasking-ai

Abstract

Artificial intelligence is reshaping the conditions under which individuals form and maintain their sense of self. As AI tools enhance professional output, many users experience a growing disconnect between what they produce and what they believe themselves capable of producing independently. This article examines the mechanisms by which AI adoption influences identity formation, introduces the concept of AI-Induced Impostorism as a distinct psychological phenomenon, and proposes a reframing of AI as a tool for identity consolidation rather than displacement. Drawing on foundational literature in impostor phenomenon research (Clance & Imes, 1978), social comparison theory (Festinger, 1954), and emerging scholarship on technology-mediated self-evaluation, this article argues that the primary risk of AI is not technological dependence, but the erosion of perceived ownership over one’s own cognitive contributions.

Keywords: impostor phenomenon, AI-induced impostorism, identity disruption, social comparison, self-efficacy, AI and confidence

1. Introduction

The rapid integration of artificial intelligence into professional, educational, and creative domains has introduced a new variable into the psychology of self-perception. Unlike previous technological disruptions, AI does not merely alter how tasks are completed. It alters how individuals interpret their own role in completing them. When a professional’s output is enhanced by an AI tool, the resulting product may exceed what that individual could produce independently. This gap between AI-assisted performance and self-assessed competence creates conditions structurally similar to those described in impostor syndrome literature.

Clance and Imes (1978) originally defined the impostor phenomenon as an internal experience of intellectual phoniness among high-achieving individuals who, despite objective evidence of success, remain convinced that they are not genuinely intelligent and fear being exposed as frauds (p. 241). Subsequent research expanded this definition beyond high achievers, identifying impostorism as a context-sensitive response to perceived performance-identity misalignment (Sakulku & Alexander, 2011). This article proposes that AI adoption constitutes a new and measurable context for such misalignment.

2. Identity in the Age of Algorithmic Mediation

Identity is not a static attribute. It is constructed through ongoing cycles of action, feedback, and interpretation (Burke & Stets, 2009). Artificial intelligence disrupts this cycle at the feedback stage. Algorithmic systems curate, filter, and amplify external signals, producing a distorted feedback environment that the individual interprets as social reality. Social media platforms, recommendation engines, and AI-generated content collectively surface an idealized representation of peers, competitors, and achievers. The individual then applies these representations as a reference point for self-evaluation.

This process aligns with Festinger’s (1954) social comparison theory, which proposes that individuals evaluate their own abilities and opinions by comparing them to others. However, Festinger’s framework was developed prior to the existence of algorithmically curated comparison pools. Contemporary AI systems do not present a random or representative sample of others. They present an optimized sample, biased toward high performance, high visibility, and high engagement. The psychological consequence is an upward comparison spiral in which internal standards shift faster than self-concept can adapt.

“When one’s internal standard shifts faster than one’s self-understanding, the result is not growth — it is confusion.” — Atkinson, 2025

3. Validation Loop: A Proposed Mechanism

The present article proposes a cyclical mechanism, termed the AI Validation Loop, to describe how AI-enhanced performance can paradoxically reduce self-confidence over time. The loop operates as follows: an individual uses AI tools to enhance the quality of their output. The enhanced output receives positive external feedback, increasing social validation. However, the individual attributes this validation to the AI’s contribution rather than their own judgment, direction, or taste. As a result, internal belief in one’s own competence does not increase proportionally with external recognition. Over time, the gap between perceived external success and internal self-attribution widens, producing the defining experience of impostorism: the sense that one does not deserve the recognition received.

This mechanism is distinct from classical impostor phenomenon in one critical respect. Traditional impostorism involves doubting one’s qualifications relative to peers. AI-Induced Impostorism involves doubting one’s ownership of one’s own work, regardless of peer comparison. The question shifts from ‘Am I as capable as they are?’ to ‘Is this truly mine?’ This distinction has significant implications for intervention design, as strategies targeting competence-based impostorism may be insufficient for ownership-based impostorism.

4. AI as a Mirror, Not a Replacement

A central argument of this article is that AI does not generate identity erosion from an external source. Rather, it reveals pre-existing vulnerabilities in identity formation. AI systems do not originate the content they produce. They respond to user input, reflect user priorities, and execute within frameworks the user provides. This means that AI outputs carry a higher degree of authorial attribution than is commonly acknowledged by users experiencing impostorism. The AI amplifies; it does not replace.

This framing is consistent with Bandura’s (1986) concept of self-efficacy, defined as an individual’s belief in their capacity to execute behaviors necessary to produce specific outcomes. Self-efficacy is domain-specific and experience-dependent. When AI handles execution, it may reduce the experiential basis on which self-efficacy is built. Awareness of this dynamic, however, can reframe AI use as a domain in which a distinct form of competence is being developed, including the ability to direct, evaluate, and improve AI-generated outputs, which are genuinely human skills.

5. Implications for Practice

The findings and framework presented here suggest several practical directions for individuals and organizations. First, intentional attribution practices may mitigate AI-Induced Impostorism. Individuals who regularly identify their own contributions to AI-assisted work, including the judgment, direction, and evaluation they provide, are more likely to maintain accurate self-attribution and sustain self-efficacy over time.

Second, community-based identity reinforcement remains essential. As Boss (1999) argued in the context of ambiguous loss, identity disruptions are most effectively addressed in relational contexts. Professional and peer communities that provide perspective, shared experience, and validation grounded in direct observation of behavior, rather than in output quality alone, offer a corrective to the comparison distortions produced by algorithmic environments.

Third, the AI Validation Loop suggests that educational and organizational interventions should address ownership cognition explicitly, helping individuals distinguish between AI-enhanced performance and AI-originated performance. This distinction is both accurate and psychologically protective.

6.Conclusion

Artificial intelligence is not taking identity. It is intensifying the conditions under which identity must be defined. The real risk is not technological dependence. It is the passive acceptance of external performance metrics as a substitute for internal self-knowledge. Individuals who fail to define their own competence clearly are more vulnerable to the distortions produced by AI-enhanced comparison environments.

The Atkinson Institute’s ongoing Identity & Resilience Study seeks to quantify these dynamics across professional and military-adjacent populations. As this research develops, the concept of AI-Induced Impostorism will be refined through empirical data. The present article establishes the conceptual framework from which that measurement proceeds.

References

Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Prentice-Hall.

Boss, P. (1999). Ambiguous loss: Learning to live with unresolved grief. Harvard University Press.

Burke, P. J., & Stets, J. E. (2009). Identity theory. Oxford University Press.

Clance, P. R., & Imes, S. A. (1978). The impostor phenomenon in high achieving women: Dynamics and therapeutic intervention. Psychotherapy: Theory, Research & Practice, 15(3), 241–247.

Festinger, L. (1954). A theory of social comparison processes. Human Relations, 7(2), 117–140.

Sakulku, J., & Alexander, J. (2011). The impostor phenomenon. International Journal of Behavioral Science, 6(1), 75–97.