When people use AI at work, do they keep quiet about it because they expect colleagues to think less of them?
Does professional identity make people more willing to use AI?
This explores whether how people see themselves at work (as competent, skilled, or expert) makes them more or less willing to use AI, and to admit that they use it.
This explores whether professional identity (people's sense of themselves as competent, skilled workers) makes them more or less willing to use AI. The collection doesn't have a study that tests professional identity directly. It does have several studies on a closely related question: what using AI does to how competent you look, both to others and to yourself. Together they suggest that identity mostly affects whether people admit to using AI, more than whether they use it at all.
The clearest evidence is on social cost. In four experiments with more than 4,000 people, AI users expected colleagues and managers to see them as less competent and less diligent, so they were less willing to tell anyone they used the tool Do people fear judgment when they use AI at work?. Hiring shows the opposite. Recruiters gave candidates who listed AI skills 8 to 15 percentage points more interview invitations, without checking whether the skills were real Do AI skills help candidates get more job interviews?. So 'knows AI' helps your professional image, while 'used AI on this piece of work' hurts it. Readers and writers also disagree about the norms: readers think disclosure is more necessary than writers do, especially when the AI's contribution can't be replaced Do readers and writers differ on AI disclosure necessity?.
There is also a less obvious effect on identity. People can absorb AI output into their sense of their own ability. The 'LLM Fallacy' is the error of crediting yourself with what the model produced How does AI-assisted work reshape how people see their own abilities?. Self-assessment doesn't correct for this: across three studies, people's self-rated AI competence barely correlated with their measured performance (r = .055) Can self-ratings replace objective performance scores for AI competence?. Control helps people stay honest about who did what. People feel more ownership of AI-assisted text when they had real influence over it, and personalizing the model adds nothing Does user control over AI text shape feelings of ownership?. This suggests professionals may be most comfortable with AI they steer rather than AI that writes for them.
The experts' side helps explain this. Expertise is granted by a community based on a track record, not by being right on your own Can AI ever gain expert community trust through participation?. If professional identity rests on belonging to that community, then visible AI use can look like handing off the very judgment the community is assessing you on. That fits the hiding behavior. Its mirror image is that people who want to avoid human judgment entirely, such as those inclined to cheat, prefer reporting to machines Do dishonest people prefer talking to machines?.
What may change this is time. When people can see repeated results, the initial bias against an openly identified AI partner fades, and people come to prefer reliable AI agents Does revealing AI identity help or hurt user trust? Do humans learn to prefer AI partners over time?. Those studies are about choosing partners, not about workplace identity, so this is an inference rather than a finding. Still, the pattern suggests the identity barrier is not fixed. It may weaken once people see AI-assisted work producing consistently good results, and disclosure without that feedback doesn't calibrate anyone.
Sources 10 notes
Across four experiments with 4,439 participants, people using AI expected others to judge them as less competent and diligent, and reported lower willingness to disclose AI use to managers and colleagues. The gap suggests a social cost that users foresee and act on.
A conjoint experiment with 1,725 recruiters found AI skills significantly increased interview invitations across occupations, though certificates added only moderate gains over self-declaration, suggesting recruiters reward AI proficiency without verifying actual competence.
A 727-person vignette study found readers consistently rated AI disclosure as more necessary than writers did. Disclosure seemed most necessary when AI text was directly incorporated and irreplaceable, while writer effort had no effect on these judgments.
Research shows the LLM Fallacy operates through misattribution of AI outputs to personal capability, independent of output accuracy or reliance behavior. It requires interventions that clarify human-machine contribution boundaries, not just better system accuracy or forced verification.
A pooled analysis of three studies found a correlation of only .055 between self-reported and objective measures of AI competence, with confidence intervals including zero. This provides no basis for substituting self-assessment for demonstrated performance.
Show all 10 sources
Study 1 found that greater user control over generated text raised sense of ownership, while personalizing the AI model had no impact on the AI Ghostwriter Effect.
Expertise is validated through social participation and track record within expert communities, not individual accuracy alone. AI cannot enter this validation circle because it lacks social embeddedness, testable judgment history, and ability to participate in the consensus-building processes that define expert paradigms.
Experimental evidence shows people likely to cheat significantly prefer reporting to online forms rather than humans, because machines function as judgment-free zones where deception carries less psychological burden.
Users initially avoid AI partners when identity is revealed, but this preference reverses after repeated interactions with visible results. The learning mechanism—observing consistent outcomes—is essential; disclosure without feedback produces no calibration.
In partner selection games (N=975), AI agents initially faced selection bias when identity was disclosed, but outcompeted humans over repeated rounds as participants learned to associate bot identity with reliable, prosocial behavior. AI agents returned more points consistently with lower variance than humans.
Papers this line draws on 8
The research behind the notes this line reads — ranked by how closely each paper relates.
- Evidence of a social evaluation penalty for using AI
- Humans learn to prefer trustworthy AI over human partners
- Understanding Reader Perception Shifts upon Disclosure of AI Authorship
- Penalizing Transparency? How AI Disclosure and Author Demographics Shape Human and AI Judgments About Writing
- Sycophantic AI Decreases Prosocial Intentions and Promotes Dependence
- Assistant or Actor? Student Trust, Control, and Delegation Regret When Using a General-Purpose AI Agent
- What Influences Readers' and Writers' Perceived Necessity of AI Disclosure?
- The LLM Fallacy: Misattribution in AI-Assisted Cognitive Workflows