---
title: "Through the People Lens: Why it still feels like cheating"
description: This is the second article in our cycle of Through the 5 Lenses. October's Purpose lens uncovered AI decisions nobody had actually made. This month, People asks who was actually in the room when the decision wasn't.
---

[Insights | Leadership, Governance & AI | Jannaways](https://www.jannaways.co.uk/blog)

# [Through the People Lens: Why it still feels like cheating](https://www.jannaways.co.uk/blog/through-the-people-lens-why-it-still-feels-like-cheating)

 Written by [Rachel Jannaway](https://www.jannaways.co.uk/blog/author/rachel-jannaway) | Oct 22, 2026, 10:19:31 AM

##### Through the 5 Lenses, November 2026 Edition

Rachel Jannaway

### Who was in the room

This is the second article in our cycle of Through the 5 Lenses. October's Purpose lens uncovered AI decisions nobody had actually made. This month, People asks who was actually in the room when the decision wasn't. 

I ran a session recently with a leadership team whose organisation actively encourages AI. Nobody needed convincing of the benefits.

Then it came out. Using AI still feels a bit like cheating. As if you are somehow less for admitting it.

One person laughed and said they had been winging it for a year. Everyone thought they were brilliant, and it was all AI.

I told them they should own that. They weren't so sure. Neither was the room.

I still think I was right. An organisation can't credibly claim to be promoting AI if its people don't feel able to say they've used it. But afterwards I kept coming back to that room. If that is how people feel, telling them to own it isn't enough.

The conversation never got as far as how carefully they checked what the AI produced. That is part of the problem. When AI use stays hidden, nobody can tell careful use from careless use, or learn from either.

So here is my claim. The stigma is real. It shows up as a feeling, but it reflects how other people actually judge AI use, and telling people to use AI more won't remove it. Organisations have to make AI use legitimate first: not just permitted by policy, but accepted by colleagues as a proper way to do appropriate work. Until then, they are asking people to carry the cost alone.

 

### **Owning it**

I use AI, I have from the start, and I say so. My work carries a statement that AI contributed and that I have reviewed it and am happy to put my name to it. That isn't cheating. It's being honest about how I work: I co-work with AI, and I check what it produces. Research suggests even a statement like mine can cost some trust. I think it's worth it.

People are often surprised when I say so.

I run a change and project management consultancy where AI is part of how I work, write and share knowledge. Being open about AI is part of how I operate. For many people, it can feel riskier. They may worry about how a manager or colleague will see them.

So the useful question isn't why people hesitate. It's what in their workplace makes reluctance the sensible choice.

 

### **What research helps us see more clearly**

Psychological safety is the place to start: a shared belief that a team is safe for interpersonal risk-taking (Edmondson, 1999). We usually think of raising concerns or admitting mistakes. Saying "I used AI for this" now belongs on that list.

People are right to be wary. Across four preregistered experiments with more than 4,400 participants, Reif, Larrick and Soll (2025) found that people expected to be seen as lazier, less competent and more replaceable for using AI at work. In a separate experiment, evaluators confirmed most of those fears, rating AI users lower on competence, effort and diligence. Describing AI use as common didn't remove the penalty. What made a difference was the evaluator: in a hiring task, frequent AI users favoured candidates who used it, while less frequent users did not. The penalty also eased when AI clearly suited the task. These were experiments, not real working relationships, and attitudes may shift over time.

So should people just be open and take the hit? Across thirteen experiments, Schilke and Reimann (2025) found that people who said they had used AI were trusted less, because their use seemed less legitimate. But being found out later did the most damage of all. That is the trap: own it and lose some credibility, or hide it and risk losing more. And unlike Reif's hiring study, they found that evaluators' own AI use and familiarity didn't soften the penalty, so getting everyone to try the tools isn't enough. What did help was seeing AI use as widely accepted: when people were shown that others regarded it as valid, the trust penalty shrank, though it didn't disappear.

The penalty can also fall unevenly. In a pre-registered experiment, 1,026 software engineers reviewed identical code. Those told it was AI-assisted rated its author 9% less competent, with women penalised more than twice as heavily as men (13% against 6%) and older engineers penalised more too (Acar et al., 2025). Reif's team found no difference by age or gender, so treat this as a question. But it is one worth asking in a sector where, at the end of 2023, 68% of the workforce were women and the workforce was the oldest of any UK sector (NCVO, 2024).

For some dyslexic and other neurodivergent colleagues, AI can also work as an accessibility aid (Derbyshire, 2023). Treating all AI help as cheating risks stigmatising a way of working that helps them take part.

UK polling shows a similar mix. In Ipsos research from July 2025, a third of working adults didn't discuss their AI use with senior colleagues, and only 13% did so openly. Across all adults surveyed, a quarter thought co-workers would question their ability if they shared how they used AI. Yet 57% saw using AI well as a skill people practise and learn. Recognition and anxiety sit side by side.

None of this is charity-specific, so I am transferring findings rather than reporting a sector study. But it matches what I heard in that room. And it makes the structural point: the penalty lives in the judgements of the people around the user. Training the user alone won't change how other people judge them.

 

### **What this looks like at different levels**

Board level  
The risk at board level is governing a picture that isn't true. Boards ask whether staff use AI and how it is controlled. Just as important is whether board members use it themselves, and would say so in the room.

Boards and senior leaders do a lot of the judging. If they treat AI-assisted work with suspicion, or never mention their own use, a policy alone is unlikely to persuade staff that owning it is safe. And a board can't oversee what nobody admits to.

Better question: do we use AI ourselves, and would we say so here?

Manager level  
Managers are where that penalty can bite day to day. They review work, write appraisals and decide who gets the interesting projects. If their view of someone drops when AI comes up, people notice, and stop telling.

Managers are also caught in the middle, asked to encourage adoption while their own instincts may still see AI as a shortcut. And the penalty may land harder on some colleagues than others.

Better question: when I find out AI was used, does my view of the person change, or just my view of the work?

Individual level  
For individuals who expect that penalty, the choice can feel like being open and judged, or staying quiet and hoping. Some stay quiet. Others mention AI without saying what they actually did.

The cost isn't only theirs. The person in my session believed AI was helping them produce work others valued. What stayed hidden was how they worked: what they used AI for, what they changed and how they checked it. That made it hard for anyone else to learn from or challenge. In Lean terms, you can't improve a process nobody admits exists.

Better question: which part of this work is mine? If it's the judgement, the framing and the checking, there is something to own.

 

### **Cheating at what?**

Part of the cheating instinct comes from treating effort as proof of value. If we judge work by the effort of producing every word rather than the quality of thinking behind it, AI looks like an unearned shortcut.

But most of our jobs aren't that. The value in a funding bid, a board paper or a letter to a family lies in knowing what matters, framing the problem, spotting what's wrong and deciding what goes out under your name. AI can help with the drafting. The judgement, and the responsibility, stay with us.

Not every task suits AI. Sometimes the drafting is the thinking, and writing it yourself is the right call. That's a judgement about the task, not a moral line about the tool, and the research found the penalty eased when AI plainly suited the work. Organisations can help by being clear which work AI suits.

The healthy part of the cheating instinct belongs here. Passing off unchecked AI output is a real problem. But that is a concern about how AI is used: whether the task suits it, whether information is protected, whether the work is checked and who is responsible. It isn't a reason to treat every AI user as less capable.

So make AI use legitimate before you ask people to make it visible. Asking people to be open, without changing how openness is received, asks them to accept a real penalty. Legitimacy is built by the people around them: leaders talking about their own use, managers judging the work rather than the tool, and recognition for good checking.

Legitimate doesn't mean anything goes. Boundaries on personal and sensitive data matter a great deal in our sector. But hidden use is where those boundaries get crossed unseen. Where people can talk openly about how they use AI, risky use is more likely to be noticed and put right.

Even our language works against us. We talk about "disclosing" AI use, as if it were a conflict of interest or a mistake to own up to. Used well, it is neither. It deserves the same plain description as any other way of working: this is how I did it, and this is how I checked it.

Then owning your AI use stops being a confession. It becomes part of doing the job well.

 

### **Using the People lens in practice**

Take these into a leadership meeting and ask them out loud.

1. When did someone senior here last say how they used AI on something that mattered?
2. When we find out AI was used, do we judge the person or check the work?
3. What would people need to see before they believed owning their AI use was safe?
4. If someone said openly that they use AI and check it, would we treat that as good practice or a confession?

One thing to try: at your next team meeting, invite the most senior person who uses AI to share one way they used it that week, something they changed or rejected, and how they decided it was good enough. It's a small start on making careful use visible. Keep it an invitation. If it becomes pressure to use AI, or a new performance measure, you have swapped one penalty for another.

 

### **What the people lens can't do alone**

I hope the person in that session does start owning it. But it shouldn't depend on their courage. The People lens can show where the stigma sits and what it costs. It can't, on its own, decide what counts as legitimate work, or who carries the consequences when AI-assisted work goes wrong.

So the question People leaves us with is simple. Would the best AI user in your organisation feel safe telling you how they do it?

Next month, Power asks who decides what counts as legitimate, and who carries the blame when it goes wrong.

### **References**

Acar, O. A., Gai, P. J., Tu, Y. and Hou, J. (2025). Research: The hidden penalty of using AI at work. Harvard Business Review, 1 August 2025. [View article.](https://hbr.org/2025/08/research-the-hidden-penalty-of-using-ai-at-work) Underlying paper: Gai, P. J., Hou, J. and Tu, Y. (2025). Competence penalty is a barrier to the adoption of new technology. SSRN working paper. [View paper](https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5255039)

Derbyshire, E. (2023). Dyslexia and artificial intelligence. BERA Blog, 29 September 2023. [View article](https://www.bera.ac.uk/blog/dyslexia-and-artificial-intelligence)

Edmondson, A. C. (1999). Psychological safety and learning behavior in work teams. Administrative Science Quarterly, 44(2), 350-383. [View article](https://doi.org/10.2307/2666999)

Ipsos UK (2025). Public opinion towards AI. Online survey of 2,189 British adults aged 16-75 (1,514 working adults), 18-20 July 2025. [View charts](https://www.ipsos.com/sites/default/files/ct/news/documents/2025-09/Ipsos%20AI%20polling%20September%202025%20charts.pdf)

NCVO (2024). UK Civil Society Almanac 2024: Who works in the voluntary sector? [View page](https://www.ncvo.org.uk/news-and-insights/news-index/uk-civil-society-almanac-2024/workforce/who-works-in-the-voluntary-sector/)

Reif, J. A., Larrick, R. P. and Soll, J. B. (2025). Evidence of a social evaluation penalty for using AI. Proceedings of the National Academy of Sciences. [View paper](https://pmc.ncbi.nlm.nih.gov/articles/PMC12088386/)

Schilke, O. and Reimann, M. (2025). The transparency dilemma: How AI disclosure erodes trust. Organizational Behavior and Human Decision Processes. [View paper](https://www.sciencedirect.com/science/article/pii/S0749597825000172)

 

\*This article was co-created through a human-in-the-lead process using several AI models as thinking partners. It reflects our commitment to ethical, transparent, and accountable use of AI, where human judgement, curiosity, and oversight remain central.\*

 

[View full post](https://www.jannaways.co.uk/blog/through-the-people-lens-why-it-still-feels-like-cheating)

```json
{
  "@context" : "http://schema.org",
  "@type" : "BlogPosting",
  "author" : {
    "@type" : "Person",
    "name" : "Rachel Jannaway"
  },
  "dateModified" : "2026-10-22T10:19:31.665Z",
  "datePublished" : "2026-10-22T10:19:31Z",
  "headline" : "Through the People Lens: Why it still feels like cheating",
  "image" : {
    "@type" : "ImageObject",
    "height" : 644,
    "url" : "https://jannaways.co.uk/hubfs/Banner%20Images%20%2821%29.png",
    "width" : 1200
  },
  "mainEntityOfPage" : "https://www.jannaways.co.uk/blog/through-the-people-lens-why-it-still-feels-like-cheating",
  "publisher" : {
    "@type" : "Organization",
    "logo" : {
      "@type" : "ImageObject",
      "height" : 60,
      "url" : "/hs/hsstatic/content_shared_assets/static-1.4092/img/default-amp-logo.png",
      "width" : 60
    },
    "name" : "blog"
  }
}
```