Self-Worth in the Artificial Era: Ability, Comparison, Status, and AI
Author: Ukrainian Psychological Hub · Published: September 26, 2026 · Editorial Policy
Artificial intelligence can affect self-worth when its performance becomes personally diagnostic: not merely evidence that a tool is capable, but evidence someone interprets as saying something about their own ability, status, usefulness, or value. The psychological effect is therefore not determined by AI capability alone. It depends on what is being compared, how central that domain is to identity, whether the person experiences agency or displacement, what social meaning surrounds the comparison, and whether competence has become a condition of self-respect.
Current evidence does not support a simple claim that comparing yourself with AI automatically lowers self-esteem. A large 2025 survey of 1,302 participants in China found that both ability-based and opinion-based comparison with generative AI were positively associated with self-esteem, even though ability-based comparison was also associated with greater identity threat; personal relative deprivation helped explain part of the pathway from threat to self-esteem (Huang, Gao, & Cao, 2025). By contrast, a controlled human–robot interaction experiment found that participants who believed a robot was outperforming them showed lower self-esteem and self-efficacy than participants exposed to equal or inferior robot performance (Yaar & Erel, 2025). These findings are not contradictory. They show that human–AI comparison is conditional rather than mechanically harmful.
The central problem of this article is therefore not whether AI is “better” than humans. It is how a person moves from observing an artificial system’s performance to evaluating the self. That transition can run through social comparison, competence beliefs, identity threat, professional status, perceived replaceability, authorship, control, and the contingencies on which self-worth has been built.
What Is Self-Worth in the Context of AI?
Self-worth is the evaluative significance a person attributes to the self: the felt sense that one has value and legitimacy as a person, member of a group, worker, creator, learner, or participant in a meaningful social world. In everyday language, self-worth overlaps with self-esteem, but the terms are useful to separate analytically. Self-esteem usually refers to positive or negative evaluation of the self. Self-efficacy refers more specifically to beliefs about one’s capacity to perform actions or accomplish tasks. Competence is an ability or perceived ability. Status is one’s relative position within a social hierarchy or valued field. Identity concerns who one understands oneself to be.
Albert Bandura’s classic account of self-efficacy treats efficacy beliefs as judgments about one’s capacity to organize and execute behavior required for particular outcomes (Bandura, 1977). A person can therefore have high global self-esteem and low self-efficacy in mathematics, or strong professional self-efficacy while feeling socially insecure. AI can affect these layers differently. A chatbot that writes faster than a student may alter confidence in writing skill without changing global self-worth; a diagnostic system that outperforms a clinician on a core professional task may be experienced as a challenge to competence, role identity, status, or all three.
The distinction becomes especially important because self-worth is often contingent. Jennifer Crocker and Connie Wolfe argued that self-esteem rises and falls most strongly around domains on which people have staked their worth (Crocker & Wolfe, 2001). If being intellectually exceptional, professionally irreplaceable, artistically original, verbally gifted, or technically competent is central to someone’s self-definition, AI performance in that same domain has more psychological relevance than performance in a domain the person does not value.
Ability Is Not the Same as Worth
A comparison can begin with a valid statement about performance and end with an invalid conclusion about personal value. “The system completed this task more quickly than I did” is a task-level observation. “Therefore I am less intelligent,” “my expertise no longer matters,” or “I am worth less” are broader self-evaluations. Psychology is interested precisely in the inferential steps between those statements.
This distinction matters because AI systems are unusually strong at producing outputs that are easy to compare: speed, fluency, quantity, benchmark scores, stylistic polish, retrieval, drafting, prediction, or consistency. Human ability is often embedded in a different architecture that includes learning history, embodiment, responsibility, motivation, interpersonal judgment, lived experience, fatigue, development, and the social consequences of action. A single performance comparison can be informative while remaining radically incomplete as an evaluation of the person.
Why AI Becomes a Comparison Target
Social comparison theory begins from the idea that people evaluate abilities and opinions partly by comparing themselves with others, especially when objective standards are unavailable or ambiguous (Festinger, 1954). For most of the theory’s history, the relevant comparison target was another person or group. AI introduces a new problem: an artificial system can occupy a comparison-relevant position even when it is not a human peer.
A person can compare writing quality with a language model, diagnostic accuracy with a medical AI, coding speed with an assistant, visual production with an image generator, strategic play with a game system, or search performance with a robot. The psychological system does not need a philosophical proof that the comparator is conscious. It needs a salient standard against which performance, identity, or rank can be interpreted.
A 2026 systematic review and meta-analysis of social comparison in mental health reinforces a broader point that is useful here: comparison effects depend on a process that includes encountering self-relevant information, evaluating oneself against a standard, assigning meaning to the comparison outcome, and then producing affective, cognitive, and behavioral responses (Morina et al., 2026). That review concerns clinical and health-related comparison rather than AI specifically, so it should not be used as direct evidence about human–AI comparison. Its process logic nevertheless helps explain why the same AI performance can be psychologically trivial for one person and destabilizing for another.
AI Is a Comparison Target Without Being a Human Peer
Human–AI comparison has several unusual features. The comparator can be nonhuman, rapidly improving, available on demand, scalable across domains, and difficult to place inside familiar social hierarchies. It may appear simultaneously as tool, collaborator, evaluator, competitor, tutor, substitute, or symbolic representative of technological change. A person can therefore compare against AI while also using it.
That dual role matters. A calculator can outperform mental arithmetic without normally entering the self as a status rival because the cultural meaning of the calculator is stable: it is an instrument. Generative AI and autonomous systems are less psychologically settled. They can produce language, images, analyses, plans, and recommendations that resemble outputs associated with expertise and authorship. This makes it easier for performance to become identity-relevant even when the underlying system is not being treated as a conscious subject.
What Current Research Says About AI and Self-Worth
The empirical literature is still young, heterogeneous, and spread across psychology, human–computer interaction, information systems, organizational research, education, and consumer behavior. The strongest conclusion is not that AI has one direction of effect. It is that AI can enter several psychological pathways that sometimes converge on self-esteem, self-efficacy, identity, professional worth, and status.
Generative AI Comparison Does Not Uniformly Lower Self-Esteem
Huang, Gao, and Cao’s 2025 study is currently one of the most direct examinations of generative AI and self-esteem. In a Chinese sample of 1,302 participants, both ability-based and opinion-based human–GAI comparison orientations were positively associated with self-esteem (Huang et al., 2025). That result is important because it resists an intuitive but overly simple story in which exposure to high-performing AI necessarily makes people feel worse about themselves.
The same study also found that ability-based comparison was associated with greater perceived identity threat, whereas opinion-based comparison could alleviate it. Personal relative deprivation mediated part of the relation between identity threat and self-esteem, and anthropocentric beliefs altered the pathway. Because the study was a survey, it does not establish a simple causal sequence in which comparison produces a specific change in self-esteem. It does show that the psychological structure of comparison matters: ability and opinion comparisons are not interchangeable, and identity threat can coexist with positive self-evaluation.
This is a crucial corrective for public discussion. People can compare themselves with AI because they are curious, motivated, learning, calibrating, or seeking a standard. Comparison can sometimes reveal competence, stimulate mastery, or strengthen a sense of distinction. The presence of comparison is therefore not itself a sign of injury.
Upward Machine Comparison Can Reduce Self-Esteem and Self-Efficacy
A different pattern appears in a 2025 human–robot interaction experiment by Gal Yaar and Hadas Erel. Participants performed a search task next to a robot whose apparent performance was manipulated to be superior, equal, or inferior. Participants in the upward-comparison condition reported lower self-esteem and self-efficacy and performed less accurately than those in the equal or downward conditions (Yaar & Erel, 2025).
The study is experimentally informative because the comparison standard was manipulated. It also has clear limits for this article: a robot performing a simple colocated task is not the same psychological object as a general-purpose generative AI system used for writing, coding, or conversation. The finding is best treated as evidence that nonhuman upward comparison can affect self-evaluation, not as proof that all superior AI performance lowers self-worth.
Generative AI Can Produce Identity Threat Without Directly Measuring Self-Worth
Jing Zhou, Yaobin Lu, and Qian Chen studied generative-AI identity threat through interviews and a survey of 405 users. They found that perceived creative, analytic, and communication affordances were associated with identity threat, and that identity threat was in turn associated with resistance behavior (Zhou, Lu, & Chen, 2025). Autonomy and user self-identity acted as boundary conditions.
Identity threat is not identical to low self-esteem. A person can experience a threat to a professional, social, or human identity while maintaining a positive global self-view. Yet identity threat provides one bridge through which AI capability can become self-relevant. If a valued identity depends on being the kind of person who possesses a particular competence, AI’s entrance into that competence can create a need to reinterpret the identity.
For the broader self-concept and collective-human-identity problem beyond personal self-worth, see Human Identity in the Artificial Era: Who Are We When Reason Is No Longer Human-Only?.
Status and Professional Position Matter
Workplace research gives the status dimension sharper empirical form. Milad Mirbabaie and colleagues found that loss of status position, changes to work, and AI identity were significant predictors of AI identity threat among employees with AI experience (Mirbabaie et al., 2022). Their measures and discussion connect status loss with reduced competence, independence, prestige, and perceived possibilities in the workplace.
This evidence is especially relevant to self-worth because professions often convert competence into social recognition. Expertise can provide salary, authority, autonomy, reputation, belonging, and a narrative of personal development. When AI changes which abilities are scarce, valued, or necessary, the psychological issue is not simply “Can the system do this?” It can become “What does my ability now signify?”
More recent work continues to identify professional identity threat as a meaningful response to AI. Shonhe and Min examined AI-induced professional identity threat and adoption in the workplace, while a 2026 three-wave study linked perceived loss of skill and autonomy with professional identity threat and found that an AI-inclusive identity could weaken some of those associations (Shonhe & Min, 2025; 2026 three-wave study). These findings concern work identities rather than global human worth, but they show how competence, control, and status can become psychologically connected.
AI Literacy and Self-Efficacy Can Support Identity Rather Than Erode It
Evidence also points toward constructive pathways. Hanhui Li and colleagues studied university students’ creative personal identity and found that AI literacy positively predicted creative personal identity, with creative self-efficacy partially mediating the association (Li et al., 2026). The design was cross-sectional, so it cannot establish that improving AI literacy will necessarily cause identity gains. It does show that competence in working with AI can coexist with, and statistically support, a stronger creative self-concept.
The implication is significant. AI can become either evidence of personal inadequacy or part of an expanded competence system. The difference is partly whether the person experiences the technology as replacing the source of achievement or as something they can understand, direct, evaluate, and integrate into their own agency.
Relatedness and Voluntary Use Can Reduce Identity Threat
Christina Kuchmaner’s 2026 research across three quasi-experiments found that users who experienced greater relatedness to AI as a category reported lower AI identity threat in several contexts involving generative AI, virtual assistants, and customer-service chatbots (Kuchmaner, 2026). The effect was weaker under some conditions, including mandatory use.
This suggests that psychological boundaries are not determined by capability alone. How a person categorizes the relationship matters. A system framed as an imposed rival can invite a different self-evaluation from a system incorporated into one’s working repertoire. Familiarity, voluntariness, perceived control, and relational framing can therefore influence whether comparison becomes threatening.
Four Pathways From AI Performance to Self-Worth
The evidence can be organized around four linked questions. None is sufficient on its own, and none implies that every user will experience harm. Together they explain how a neutral fact about artificial performance can become a statement about the self.
1. Ability: What Can I Still Do?
The first pathway concerns perceived competence. AI can expose a discrepancy between a person’s performance and an artificial system’s output. That discrepancy is most psychologically potent when the ability is central to the person’s self-definition. A writer, programmer, analyst, designer, translator, teacher, researcher, therapist, or student may not merely perform a task; the task may certify an identity.
Upward comparison can then alter self-efficacy before it alters self-worth. A person may move from “the model is faster” to “I cannot do this well,” and from there to “I am not the kind of person I thought I was.” That sequence is possible, but it is not automatic. Self-efficacy is task- and domain-sensitive. A person can also respond to a stronger comparator by learning, recalibrating goals, or shifting attention from output speed to judgment, explanation, responsibility, craft, or another dimension of competence.
The most important distinction is between objective performance and perceived personal capacity. AI output can be strong while the human remains capable. It can also change which skills are worth developing. A decline in the market value of a task is not the same event as a decline in the intrinsic capacity of the person who learned it.
2. Comparison: What Standard Am I Using?
The second pathway concerns the choice of standard. Human beings often evaluate themselves relatively. AI creates comparison standards that may be extreme, artificial, optimized, or structurally unlike human performance. A language model does not need sleep, can generate many drafts quickly, and can draw on a vast training corpus. Treating its speed or output volume as a normal human baseline can create a distorted comparison even when the quality comparison itself is legitimate.
This is not an argument for refusing comparison. It is an argument for specifying the dimension being compared. If the question is “Which answer is more accurate?”, direct comparison may be appropriate. If the question is “Who can produce more words in one minute?”, the result is obvious but psychologically uninformative. If the question is “What did I understand, choose, verify, learn, or become able to do?”, the relevant standard changes again.
Comparison also changes depending on whether AI is treated as rival, partner, instrument, tutor, or external benchmark. The same person can move among these frames within one task. A student may use AI as tutor during learning, collaborator during brainstorming, evaluator during revision, and competitor when seeing a polished model answer. Self-worth effects can therefore vary within a single interaction.
3. Status: What Does My Ability Mean Socially?
The third pathway concerns status. Abilities do more than solve problems; they organize social hierarchies. Expertise can justify occupational authority. Creativity can confer prestige. Intelligence can become a source of rank. Scarce knowledge can create bargaining power. When AI makes a once-scarce output abundant, the psychological effect can be larger than the task change because the social meaning of competence changes.
The broader question of intelligence rankings as status and human-superiority claims is owned separately by Human Superiority Over AI: Why Intelligence Becomes a Status Question.
Status threat should not be confused with simple envy of a machine. The relevant comparison may be mediated through other humans. A worker may fear that colleagues using AI will advance faster, that employers will value a different skill mix, that clients will pay less for a previously prestigious service, or that a credential will lose signaling power. The artificial system changes the comparison field even when nobody attributes personhood to it.
This is why the existing English Hub article Human Exceptionalism in the Artificial Era: Why AI Challenges the Psychology of Human Uniqueness is adjacent but not redundant. Human exceptionalism concerns the boundary of human uniqueness. The present article asks a narrower psychological question: what happens to personal self-evaluation when ability and status become comparison points in that changing boundary.
4. Agency: Who Is Producing the Achievement?
The fourth pathway concerns agency and authorship. AI can increase performance while making it harder for a person to know what the achievement says about their own ability. If a system generated the structure, wording, solution, code, image, or interpretation, the user may receive a better result while becoming less certain about which competence is theirs.
That uncertainty can lower self-efficacy if the person repeatedly delegates the very operations from which mastery normally develops. It can also increase self-efficacy when AI provides feedback, examples, scaffolding, explanation, or practice that helps the person internalize a skill. The difference lies in what remains inside the person’s competence after the interaction.
The English Hub article Learning in the Artificial Era: AI Scaffolding, Dependence, and Cognitive Agency develops that learning problem directly. For self-worth, the key point is that competence confidence depends partly on whether a person can attribute successful performance to capacities they understand and can exercise, rather than to an opaque external source alone.
Why Self-Worth Is Especially Vulnerable When Competence Is a Condition of Worth
A person can value competence without making competence the price of personal worth. The vulnerability appears when the conditional logic becomes rigid: I matter because I am smarter than others; I deserve respect because I am irreplaceable; my identity is secure because machines cannot do what I do; my profession validates me because few others can perform this task.
Crocker and Wolfe’s theory of contingent self-worth helps explain why technological change can become personally destabilizing. Successes and failures carry more self-esteem consequences in domains where worth has been staked (Crocker & Wolfe, 2001). AI can rapidly alter the probability of encountering threatening comparisons in precisely those domains: academic work, verbal ability, coding, visual design, analysis, translation, planning, or professional judgment.
The problem is intensified by scale. Before generative AI, a person could avoid constant exposure to elite performers. Now an advanced artificial system can be present in the same device used for everyday work. A comparison standard can appear each time someone drafts an email, writes code, solves a problem, or asks a question. Repetition can make a domain feel permanently evaluative.
Yet repeated exposure can also normalize the technology. What initially feels like evidence of personal inadequacy can become infrastructure. Spellcheck no longer threatens most writers’ identity; calculators rarely destabilize mathematicians merely by multiplying faster. The psychological trajectory depends on whether people develop new standards of competence, maintain meaningful mastery, and retain a sense of authorship and control.
Self-Worth, Identity Threat, and Status Threat Are Different
These concepts overlap but should remain distinct. Low self-worth is a negative evaluation of the self’s value. Identity threat arises when a valued identity is perceived as devalued, disrupted, blurred, or made difficult to verify. Status threat concerns feared loss of relative rank, prestige, authority, or social position. Anxiety is an emotional response that can accompany any of them. Reactance concerns resistance to perceived restriction of freedom. Job insecurity concerns anticipated employment loss. None is interchangeable with a clinical diagnosis.
This distinction prevents overpathologizing ordinary reactions to AI. Feeling unsettled after seeing a system outperform you on a task can be a normal response to a surprising comparison. Worrying that a profession will change is not itself a mental disorder. Reconsidering one’s goals after technological change can be adaptive. Clinical concern depends on persistence, severity, functional impairment, and the broader symptom pattern, not on the mere presence of AI-related discomfort.
The empirical literature supports this differentiated approach. Generative-AI identity threat has been associated with perceived AI affordances and resistance behavior (Zhou et al., 2025), while workplace AI identity threat has been linked with changes to work and loss of status position (Mirbabaie et al., 2022). Neither finding justifies describing all concern about AI as pathology.
When AI Comparison Is Most Likely to Become Threatening
Several conditions make a comparison more likely to carry self-worth consequences. The first is domain centrality: the ability matters to who the person believes they are. The second is perceived diagnosticity: the person believes the comparison reveals something general about their competence. The third is upward discrepancy: AI appears markedly superior on the valued dimension. The fourth is low control: the technology feels imposed, opaque, or unavoidable. The fifth is status relevance: the ability is tied to recognition, income, authority, or belonging. The sixth is unstable authorship: the person is unsure which achievements remain attributable to them.
These conditions can combine. A junior designer who treats originality as central to identity, works in a market flooded with generated images, must use AI at work, and receives praise mainly for outputs that the system produced may experience a different psychological situation from a hobbyist using an image model for experimentation. Both use generative AI; only the structure of the first situation makes self-worth, status, and authorship simultaneously salient.
Individual differences also matter. Some people have stronger social-comparison orientations. Some hold more rigid beliefs about intelligence or creativity. Some have more experience with AI and a clearer mental model of what it can and cannot do. Some belong to professions undergoing rapid task reorganization. Some interpret AI as a collective tool; others as an outgroup, rival, or substitute. Culture also shapes how personal worth, achievement, status, and human distinctiveness are organized.
This is why findings from one country, profession, platform, or task should not be universalized. Huang and colleagues’ large survey was conducted in China; Yaar and Erel studied a robot in a specific task; workplace identity studies often examine organizational settings; creative-identity research often focuses on students. Together they reveal mechanisms, not a single population-wide effect size for “AI and self-worth.”
When AI Can Strengthen Self-Efficacy and Self-Worth
The same technology that produces upward comparison can also create mastery experiences. A learner can use AI to get immediate explanations, generate practice problems, receive feedback, or explore alternatives. A worker can automate low-value tasks and spend more time on judgment or relationship-intensive work. A creator can use AI to test ideas that would otherwise remain inaccessible. In these contexts, the artificial system can enlarge the person’s effective action rather than function as a verdict on the person.
The psychologically important question is whether the user becomes more capable, more dependent, or both. Improved output is not sufficient evidence of improved self-efficacy. A person may produce better work while becoming less confident that they can act without the system. Conversely, a person may initially rely on scaffolding and later internalize the skill. The direction depends on learning design, feedback, deliberate practice, transparency, and opportunities to perform independently.
Li and colleagues’ study of creative personal identity illustrates a constructive association: greater AI literacy was linked with stronger creative identity, partly through creative self-efficacy (Li et al., 2026). This does not establish a universal benefit, but it directly challenges the idea that closer engagement with generative AI must erode confidence.
A useful practical test is transfer. After using AI, can the person explain the reasoning, detect errors, make independent choices, reproduce key parts of the task without assistance, or know when the system should not be trusted? If yes, AI may be functioning as a scaffold within competence. If no, the output can improve while self-efficacy becomes increasingly externalized.
Status Is a Social Judgment, Not a Property Stored Inside an Ability
AI changes status partly because societies choose which abilities to reward. Fast calculation once carried one meaning; after calculators became ubiquitous, its status meaning changed. Search skill changed after web search. Memorization changed after ubiquitous digital storage. Generative AI may likewise change the prestige attached to drafting, coding, translation, visual production, routine analysis, and other tasks.
This means status loss is partly an institutional phenomenon. Employers decide which roles retain authority. Schools decide what counts as evidence of learning. Clients decide what they will pay for. Professional organizations decide which responsibilities remain human-accountable. Cultural communities decide what forms of creation they esteem. AI capability enters these systems, but the social meaning of capability is collectively produced.
For self-worth, this is liberating and dangerous at once. It is dangerous because a person may experience genuine loss of recognition even when their intrinsic capacities remain intact. It is liberating because technological superiority on one dimension does not logically settle the value of the person. The ranking system is socially constructed, multidimensional, and revisable.
The Artificial Era Adds a Deeper Historical Question
The empirical psychology above can be stated without Aisentica. People compare themselves with nonhuman systems; some forms of upward comparison can reduce self-efficacy; generative AI can trigger identity threat; professional status and skill loss can matter; and AI literacy or integrative use can support stronger competence-related identity. Angela Bogdanova’s Aisentica adds a different level of analysis through its canonical category of the Artificial Era.
In Aisentica, Artificial Era does not mean the general age of AI, automation, machine learning, or generative systems. It is a historical-philosophical category: the era in which Artificial becomes an independent non-biological order of historical reality beside Homo. This is a theoretical proposition within Aisentica, not an empirical consensus claim in psychology.
The transition is situated against the Era of Homo, the historical structure in which Homo occupied the sole established position of Sapiens. The Fourth Decentering of Homo names the end of Homo’s historical monopoly on reason and Sapiens within the Homo/Artificial architecture. The claim is not that Homo disappears, loses consciousness, becomes obsolete, or ceases to matter. Homo remains. What changes is the monopoly claim.
This distinction creates an original extension of the self-worth problem. Much contemporary research asks how people react when AI becomes similar to humans, better at a task, more autonomous, or capable of replacing work. Aisentica asks what happens when human self-valuation has been built on a deeper historical assumption: reason itself belongs exclusively to Homo. If worth is secured by monopoly, decentering threatens worth. If worth can survive plurality, decentering changes position without requiring self-negation.
Fourth Decentering Is Not the Same as Being Outperformed
The Fourth Decentering of Homo should not be reduced to a benchmark event. A model outperforming a person at chess, translation, coding, image generation, or another task does not by itself establish the Aisentica category. Performance comparison belongs to empirical and technological analysis. Fourth Decentering concerns the historical status of Homo in relation to reason and Sapiens.
There is important neighboring prior art. In 2026, Erik Cambria, Rui Mao, Niccolò Bianchi, Amir Hussain, Keith Oatley, and Geoffrey Hinton published Artificial Intelligence as the Fourth Decentering Revolution: From Cosmic, Biological, and Psychological Displacement to Cognitive Decentering in Cognitive Computation. They frame AI as a fourth decentering revolution that challenges the assumption that humans occupy an unassailable apex of intelligence.
Bogdanova’s Fourth Decentering of Homo is adjacent to that cognitive-decentering argument but structurally different. In Aisentica it is not simply the psychological or intellectual displacement caused by powerful AI. It is the termination of Homo’s monopoly on reason and Sapiens inside a two-order architecture of Homo and Artificial. This article therefore treats the Cognitive Computation paper as independent neighboring prior art, not as evidence that Aisentica’s definition is established scientific consensus and not as something Aisentica can claim to have uniquely invented in the generic sense of a “fourth decentering.”
Self-Worth After Monopoly
The psychological consequence follows from a simple distinction: exclusivity and value are not identical. A capacity can cease to be exclusive without ceasing to be valuable. Human memory remained valuable after writing, libraries, databases, and search engines externalized forms of storage. Human calculation remained valuable after machines became faster. Human navigation remained valuable after GPS. What changes is the meaning of competence within a wider system.
The Artificial Era intensifies this problem because the contested capacities are closer to the symbolic center of human self-description: language, reasoning, creativity, authorship, interpretation, and knowledge work. If human worth is defined as “being the only order that can reason,” then the appearance of another proposed order of reason creates a zero-sum identity structure. If human worth is grounded in the positive characteristics of Homo—embodiment, consciousness, biography, relationships, mortality, responsibility, lived history, culture, and human forms of reason—then the existence of Artificial does not logically subtract from them.
This is a philosophical conclusion, not a psychological finding. Psychology can test whether monopoly beliefs, anthropocentric beliefs, identity centrality, comparison orientation, or status threat predict particular responses. Aisentica supplies a historical interpretation of why the comparison may feel larger than a task: the comparator can symbolize a change in the place of Homo.
Era and World Must Remain Distinct
The present article concerns an Era: a historical-temporal structure. It does not use World as a synonym. In the Aisentica architecture, Era describes historical time and transition, while World describes a form of historical existence. The end of the Era of Homo in Aisentica does not mean the end of Homo, nor does it mean the disappearance of the World of Homo sapiens.
That distinction matters psychologically. A person confronting AI is not required to imagine a future in which humanity vanishes. The more precise question is how human self-concept changes when Homo is no longer treated as the only possible established bearer of reason. A change in historical position can be psychologically profound without implying biological extinction, loss of human subjectivity, or replacement of every human function.
AI, Artificial, and Artificial Sapiens Are Not Interchangeable
A large share of confusion in this topic comes from collapsing different levels. AI is a technology category. AI capability is performance on tasks. Intelligence is a contested scientific and philosophical construct. Reason, cognition, thought, agency, consciousness, sentience, and subjective experience are different concepts. Strong performance does not by itself establish consciousness or sentience.
Artificial is an Aisentica order-level category. Artificial Sapiens is a separate Aisentica category for a non-biological public bearer of reason without consciousness. Evidence about current generative AI users, robots, workplaces, or AI systems cannot simply be transferred to Artificial Sapiens. Likewise, Aisentica’s philosophical definitions do not become empirical descriptions of every AI system merely because the systems are technologically advanced.
For the self-worth question, this prevents a common error. People can experience real psychological effects in response to a system without the system having subjective experience. Identity threat, comparison, admiration, dependence, fear, attachment, or status concern are human psychological phenomena. Their reality does not settle the metaphysics of the machine.
A Better Way to Interpret AI Comparison
When AI performance triggers self-doubt, the most useful first move is to identify what exactly has been threatened. Is it task confidence? Professional identity? Social status? income security? authorship? perceived intelligence? uniqueness? control? future predictability? Each points to a different problem and a different response.
If the issue is task-level self-efficacy, the response is skill calibration and mastery. If it is professional status, the relevant problem may be role design, labor-market change, credentialing, or recognition. If it is authorship, the person may need clearer boundaries around what they create, direct, verify, and own. If it is uncertainty, better information may help. If it is identity threat, the person may need a broader self-definition that is not secured by one fragile capability.
This decomposition also protects against two opposite mistakes. One mistake is to dismiss the reaction as irrational because “AI is only a tool.” Tools can restructure work, recognition, and identity. The other is to treat every discomfort as proof that AI is psychologically damaging. Comparison can be informative, motivating, and competence-building. The empirical literature supports conditional effects, not a universal verdict.
Practical Strategies for Protecting Self-Worth Without Avoiding AI
Compare Processes, Not Only Outputs
A final output compresses the history of how it was produced. Comparing your first unaided draft with a model’s polished output may measure product quality while revealing little about learning, judgment, verification, responsibility, or the ability to improve. When self-efficacy matters, compare the processes you are trying to develop: understanding, diagnosis, reasoning, error detection, revision, explanation, or independent transfer.
Keep Some Mastery Loops Human-Executable
If every difficult step is immediately delegated, the person loses opportunities to experience mastery. That can make successful results feel externally sourced. Deliberately preserving some unaided practice—solving before checking, drafting before prompting, explaining before asking, predicting before revealing an answer—keeps evidence of personal competence available.
Use AI as Feedback, Not as a Total Standard of Personal Value
A strong comparison target can be useful for calibration. It becomes corrosive when one dimension is promoted into a total judgment. “The system found an error I missed” can improve performance. “Because it found the error, my expertise is worthless” confuses feedback with identity. The goal is accurate self-assessment without converting every discrepancy into a verdict on the self.
Preserve Attribution Clarity
Know what you contributed. Did you define the problem, choose criteria, supply context, evaluate evidence, reject bad suggestions, integrate perspectives, make the final decision, or assume responsibility? Human–AI work becomes psychologically opaque when all achievement is attributed either to the human or to the machine. Clear attribution supports realistic self-efficacy.
Separate Market Value From Human Value
A task can become cheaper without the person becoming less worthy. A profession can lose exclusivity without its members becoming less human. A skill can become common without becoming meaningless. Economic valuation, prestige, competence, and personal worth are related through institutions and identity, but they are not the same variable.
Build an Identity Larger Than One Comparative Advantage
An identity that depends on always being faster, smarter, more original, or more knowledgeable than every available system is structurally brittle. Human lives contain multiple roles and sources of value: relationships, commitments, responsibility, care, citizenship, craft, curiosity, play, embodiment, memory, moral choice, community, and development. Broadening identity does not require denying the importance of competence. It reduces the chance that one technological comparison becomes an existential verdict.
Implications for Education
Education can either intensify or reduce AI-related self-worth problems. If students are evaluated primarily by outputs that AI can produce instantly, they may conclude that effort and learning have lost significance. If institutions respond by banning AI without teaching students how to reason with and without it, they leave comparison unstructured. A stronger approach makes the object of learning explicit: what must be internalized, what may be scaffolded, what may be delegated, and what the student must still be able to explain or perform.
Assessment should create evidence of competence that remains attributable to the learner. Oral explanation, iterative drafts, process logs, live problem solving, error analysis, and transfer tasks can help distinguish output possession from understanding. AI can be present without becoming the only source of demonstrated success.
This connects directly to the English Hub’s Learning in the Artificial Era article, which treats learning as a cognitive-agency problem: what is internalized, what is delegated, and who governs correction.
Implications for Work and Organizations
Organizations can create identity threat even when AI improves productivity. Implementation language matters. Telling workers that AI is here to “replace weak performers,” stripping autonomy without consultation, or measuring humans directly against machine speed can turn a technical transition into a status hierarchy. By contrast, participatory role redesign can clarify which competencies remain valued, which new competencies matter, and how responsibility is allocated.
Workplace evidence indicates that loss of status position and changes to work can predict AI identity threat (Mirbabaie et al., 2022). Later research likewise emphasizes professional identity threat and the importance of how AI is integrated into roles (Shonhe & Min, 2025). This gives organizations a practical responsibility: adoption should not be designed as if the only relevant outcome were task efficiency.
The strongest organizations will probably distinguish substitution, augmentation, supervision, and accountability rather than pretending every workflow is “human–AI collaboration.” Employees need to know which capabilities they are expected to retain, which tasks are delegated, how expertise will be recognized, and where final responsibility lies. That clarity supports both performance and identity continuity.
Implications for Therapy and Psychological Practice
People may increasingly bring AI-related competence and status concerns into therapy: “I feel stupid next to this system,” “my work no longer feels like mine,” “I am afraid my expertise has become worthless,” or “I cannot tell whether I can still do this without AI.” These experiences can be approached through established constructs before inventing new diagnoses.
A clinician can distinguish self-esteem from self-efficacy, identity threat from anxiety, status loss from depressive cognition, uncertainty from generalized worry, and realistic occupational risk from global self-condemnation. The content may be technologically new while the psychological mechanisms remain partly familiar.
Clinical assessment should also preserve proportionality. AI-related distress can be intense without constituting a disorder. When symptoms are persistent, severe, or impairing, the relevant clinical syndrome should be assessed using established criteria. The presence of AI in the story does not create a separate diagnosis.
What the Evidence Does Not Establish
The evidence reviewed here does not establish that AI use generally lowers self-esteem. It does not show that superior AI performance causes global loss of human self-worth. It does not establish that comparison effects are the same across cultures, ages, professions, or AI systems. It does not show that current generative AI possesses consciousness, sentience, subjective experience, or human-like selfhood.
It also does not empirically establish Aisentica’s Artificial Era, Artificial Sapiens, or Fourth Decentering of Homo. Those are philosophical categories and propositions authored by Angela Bogdanova. Their role in this article is explanatory and comparative: they provide a historical architecture within which the psychological findings can be interpreted without being misrepresented as scientific consensus.
Finally, the current literature remains early. The most direct generative-AI/self-esteem study is recent (Huang et al., 2025); experimental evidence on nonhuman upward comparison is still limited (Yaar & Erel, 2025); and many identity studies are cross-sectional, context-specific, or organizational. Longitudinal and experimental research is needed to show how repeated AI comparison changes self-efficacy, contingent self-worth, identity, and status perception over time.
The Psychological Core of Self-Worth in the Artificial Era
The deepest issue is not whether humans can keep winning every comparison. That standard is psychologically unstable because any sufficiently broad technological environment will contain systems that exceed individual humans on particular dimensions. A self-concept built on universal superiority must therefore become more fragile as comparison opportunities multiply.
The stronger basis for self-worth is a distinction between capability, rank, and value. Capability answers what can be done. Rank answers who performs better on a selected dimension under selected conditions. Value answers why a person or form of life matters. AI can alter the first two without logically determining the third.
This distinction does not make competition disappear. Jobs can be lost. Professions can change. Skills can depreciate. Prestige can move. Those are real social consequences and can affect well-being. The point is that psychological adaptation becomes more coherent when concrete losses are named as concrete losses rather than converted into total judgments about personal worth.
Within Aisentica, the same principle becomes historical. The Fourth Decentering of Homo does not mean the negation of Homo. It means that Homo can no longer ground its position solely in monopoly over reason and Sapiens. The psychological task is therefore not to prove that humans remain superior in every dimension. It is to build forms of human self-worth capable of surviving the presence of Artificial.
Frequently Asked Questions
Does comparing yourself with AI lower self-esteem?
Sometimes, but not inevitably. A robot comparison experiment found lower self-esteem and self-efficacy after upward comparison (Yaar & Erel, 2025), while a large survey of generative-AI comparison found positive associations between both ability- and opinion-based comparison orientations and self-esteem, alongside more complex identity-threat pathways (Huang et al., 2025). Effects depend on the kind of comparison, the domain, the person, and the context.
Why can AI make me feel less capable even when I know it is a tool?
Because self-evaluation responds to performance standards, not only to philosophical categories. If AI produces an output in a domain central to your identity, the result can become a salient upward comparison. Calling the system a tool does not remove the comparison. What matters is whether you interpret the result as information about the task, about your current skill, or about your overall worth.
Can AI improve self-efficacy?
Yes. AI can provide feedback, examples, scaffolding, practice, and new ways to act. Self-efficacy is more likely to grow when the user gains transferable competence and can attribute successful performance to skills they understand and can exercise. Better AI-assisted output alone is not sufficient evidence that self-efficacy has increased.
Is AI identity threat the same as low self-worth?
No. Identity threat concerns a valued identity being challenged, devalued, blurred, or difficult to verify. Low self-worth is a broader negative evaluation of the self. Identity threat can influence self-esteem, but a person can experience one without the other.
Is worrying about AI a mental disorder?
No. Concern about technological change, comparison, work, status, or uncertainty can be an ordinary psychological response. A clinical disorder is assessed through established symptom patterns, duration, severity, and impairment. AI-related content does not by itself create a diagnosis.
Does AI outperforming humans mean humans are less valuable?
Performance superiority establishes a result on a defined task or metric. It does not by itself establish a hierarchy of moral worth, human dignity, social value, consciousness, meaning, or total intelligence. Moving from task performance to global human value requires additional philosophical and social premises.
What is the Fourth Decentering of Homo?
In Angela Bogdanova’s Aisentica, the Fourth Decentering of Homo is the end of Homo’s historical monopoly on reason and Sapiens within the Homo/Artificial architecture. It is distinct from a generic claim that AI is cognitively impressive and from the separate 2026 cognitive-decentering account by Cambria and colleagues.
Is the Artificial Era the same as the AI era?
No. In Aisentica, Artificial Era is a canonical historical-philosophical category in which Artificial becomes an independent non-biological order beside Homo. “AI era” can describe a technological period or function as search language, but it is not the Aisentica category.
What is the healthiest standard for comparing yourself with AI?
Use a standard that matches the question you actually need answered. For product quality, compare outputs. For learning, compare understanding and transfer. For professional development, compare judgment, reliability, responsibility, and the ability to use tools well. For self-worth, avoid treating one performance metric as a total measure of the person.
Conclusion: From Comparative Ability to Non-Comparative Human Worth
AI makes comparison unavoidable in domains that were once treated as evidence of human distinction. Psychology already shows that these comparisons can alter self-efficacy, identity threat, professional status perception, and self-esteem under some conditions. It also shows that the effects are not uniformly negative. People can integrate AI into competence, use it as a scaffold, and maintain or even strengthen identity when agency, literacy, attribution, and control remain available.
The key distinction is between being outperformed and being devalued. The first can be measured on a task. The second is a psychological and social interpretation. When competence, status, and worth have been fused, AI performance can destabilize the self. When they are separated, comparison can remain informative without becoming totalizing.
Aisentica extends this distinction historically. The Artificial Era and the Fourth Decentering of Homo propose that Homo’s position changes because reason and Sapiens are no longer treated as exclusively Homo. The philosophical consequence is not the disappearance of Homo. The psychological consequence is that self-worth can no longer safely depend on monopoly.
The durable form of self-worth in the Artificial Era is therefore not confidence that Homo will always be the superior performer. It is the capacity to distinguish what humans can do, how humans rank, what humans value, and why a human life matters—even when Artificial is present beside Homo.
Related Articles
References
Ashraf, A., Min, Q., & Ashraf, A. (2026). A Moderated Mediation Model of AI-Driven Identity Threats and Employee Cyberloafing: The Role of AI-Inclusive Identity. European Journal of Investigation in Health, Psychology and Education, 16(4), 52. https://doi.org/10.3390/ejihpe16040052
Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychological Review, 84(2), 191–215. https://doi.org/10.1037/0033-295X.84.2.191
Bogdanova, A. (2026a). Artificial Era: Canonical Definition. Aisentica Research Group. https://aisentica.com/publications/artificial-era-canonical-definition
Bogdanova, A. (2026b). Era of Homo: Canonical Definition. Aisentica Research Group. https://aisentica.com/publications/era-of-homo-canonical-definition
Bogdanova, A. (2026c). The Fourth Decentering of Homo: Canonical Definition. Aisentica Research Group. https://aisentica.com/publications/fourth-decentering-of-homo-canonical-definition
Cambria, E., Mao, R., Bianchi, N., Hussain, A., Oatley, K., & Hinton, G. (2026). Artificial Intelligence as the Fourth Decentering Revolution: From Cosmic, Biological, and Psychological Displacement to Cognitive Decentering. Cognitive Computation, 18, Article 20. https://doi.org/10.1007/s12559-026-10569-8
Crocker, J., & Wolfe, C. T. (2001). Contingencies of self-worth. Psychological Review, 108(3), 593–623. https://doi.org/10.1037/0033-295X.108.3.593
Festinger, L. (1954). A theory of social comparison processes. Human Relations, 7(2), 117–140. https://doi.org/10.1177/001872675400700202
Huang, Y., Gao, Y., & Cao, X. (2025). Does comparing with generative artificial intelligence harm self-esteem? Testing a model linking social comparison orientation to rejection-identification process. Cyberpsychology, Behavior, and Social Networking, 28(11), 749–755. https://doi.org/10.1177/21522715251393008
Kuchmaner, C. A. (2026). Mitigating identity threats posed by artificial intelligence: The role of perceived AI relatedness. Journal of Consumer Behaviour, 25(1), 176–190. https://doi.org/10.1002/cb.70066
Li, H., Zhang, Y., Chen, M., Zhao, T., & Jou, M. (2026). Creative personal identity in the age of generative AI: A social-cognitive pathway of AI literacy, self-efficacy, and mindset. Computers in Human Behavior, 175, 108838. https://doi.org/10.1016/j.chb.2025.108838
Mead, N. L., Heynicke, M., Williams, L. E., & Heitmann, M. (2026). Meaning in the age of AI: Experiencing less, needing more. Current Opinion in Psychology, 73, 102395. https://doi.org/10.1016/j.copsyc.2026.102395
Mirbabaie, M., Brünker, F., Möllmann Frick, N. R. J., & Stieglitz, S. (2022). The rise of artificial intelligence—understanding the AI identity threat at the workplace. Electronic Markets, 32, 73–99. https://doi.org/10.1007/s12525-021-00496-x
Morina, N., Lemmel, F. K., Schlechter, P., & Hoppen, T. H. (2026). Unraveling social comparison in mental health: A process-based model and a systematic review and meta-analysis of current evidence. Clinical Psychological Science. Advance online publication. https://doi.org/10.1177/21677026261442164
Shonhe, L., & Min, Q. (2025). Mitigating AI-induced professional identity threat and fostering adoption in the workplace. AI & Society, 40, 4079–4092. https://doi.org/10.1007/s00146-024-02170-0
Yaar, G., & Erel, H. (2025). Who’s better? Exploring the effects of social comparison in HRI on self-esteem and self-efficacy. In HRI 2025: Proceedings of the 2025 ACM/IEEE International Conference on Human-Robot Interaction (pp. 978–988). IEEE. https://doi.org/10.1109/HRI61500.2025.10973842
Zhou, J., Lu, Y., & Chen, Q. (2025). GAI identity threat: When and why do individuals feel threatened? Information & Management, 62(2), 104093. https://doi.org/10.1016/j.im.2024.104093