Resistance As An Expression Of Identity

A few months ago we wrote about how the people most resistant to AI adoption are often the most experienced professionals in the organisation. That note struck a nerve, enough to make us explore this point in more depth.

Our view is that this form of resistance can be mistaken for ignorance. When that happens it discards the decades these individuals have spent building expertise, replacing it with a sense of being devalued. Treating them as a change management problem to be overcome rather than injecting much needed empathy to what is an inherently human problem that requires a human response.

However, we are living in a world where the conversation about AI adoption is dominated by discussions around strategy, implementation, governance and change management. For a technology disrupting the human experience of work at this scale, the business conversation is almost entirely silent about the thing that actually determines whether adoption succeeds or fails.

What it will most certainly do to people’s sense of who they are.

The Research

The academic research is starting to look at the relationship between AI and resistance to adoption. A particularly interesting study was published in 2022 by Jussupow, Spohrer and Heinzl at the University of Mannheim, in JMIR Formative Research. The study examined how experienced physicians responded to AI systems capable of performing diagnostic tasks. (https://formative.jmir.org/2022/3/e28750)

It found two distinct dimensions of professional identity threat. The first was a threat to expert status, the sense that years of accumulated knowledge and skill were being rendered ordinary. The second was a threat to professional capability, the sense that the autonomous judgment that defined their role was being displaced.

While the mechanism for each was different, both dimensions predicted resistance to AI. The threat to capability directly produced resistance, while the threat to expert status worked through something the researchers called self-threat; a deeper, more fundamental challenge to the person’s sense of who they were, not just what they could do.

A more recent study broadened this lens beyond the medical profession to organisations across sectors. Published in 2025 by Golgeci, Ritala, Arslan, McKenna and Ali, in Human Resource Management Review, the study identified three deeper individual concerns underlying AI resistance across organisations;

- Mistrust of the technology,

- Existential questioning or a deeper form of self-examination about what the technology’s arrival means for their sense of purpose and professional worth, and

- Technological reflection or the process of asking what the technology’s arrival means for who I am and what I am worth.

(https://www.sciencedirect.com/science/article/pii/S1053482224000652)

The Investment-Resentment Paradox

The data from 2026 on white collar AI resistance shows that firms that spent the most on AI in 2024 and 2025 without involving workers in the design of how those tools integrate into actual workflows are now managing the highest resistance rates. The correlation is consistent enough to be structurally significant. High AI investment without worker co-design accelerates resentment rather than adoption.

When leadership makes a high-visibility AI investment such as investments in new platforms, mandatory training programmes or productivity dashboards without involving the people whose work will change, the signal sent to those people receive is; we are not investing in you but we have already decided how you will work.

The larger the investment announcement, the louder that signal becomes and the greater the fear felt by those impacted. While some might lean in to the process as a self-preservation mechanism, the data shows that the larger majority are likely to resist in an attempt to slam the brakes on change.

For the experienced professional whose identity is built around expertise, judgment and the autonomous exercise of professional capability, that signal is a direct challenge to the thing they value most about themselves. The organisation is telling them, through its actions, that their judgment about how their own work should be done does not matter. That the decades of accumulated expertise that inform that judgment are less relevant than the decision made by people who have never done the work.

MIT’s Work of the Future task force has made the causal mechanism explicit. Workers who lack agency in technology adoption decisions show measurably higher resistance even when they evaluate the technology itself as useful. Agency, not comprehension, is the operative variable. (https://workofthefuture-taskforce.mit.edu/)

For organisations navigating AI driven digital transformation, this is a dangerous finding to ignore.

The problem is not that people do not understand AI. The problem is that they do not have a genuine voice in how it enters their professional lives. And for people whose professional lives are the primary locus of their identity and self-worth, that absence of agency is a profound indignity.

The Seniority Paradox

Seniority is another aspect worth touching on, particularly as the people most likely to experience AI as an identity threat are often the most senior, the most experienced and the most accomplished within these organisations.

While our expectation might be that the most vulnerable are the lowest skilled and the most resistant are the least sophisticated, the evidence suggests a more complex situation.

The senior partner who has spent thirty years building a reputation for judgment in a specific domain has more invested in that domain than the associate who joined three years ago. The experienced consultant whose entire professional identity is organised around a particular form of expertise has more to lose from its devaluation than the graduate who is still building theirs. The seasoned analyst whose self-worth is grounded in the quality of their insight has more at stake when an AI system produces comparable insights in seconds.

Seniority, in this context, is a vulnerability, particularly as AI can become an accelerator for those who can rely on it to bridge gaps within their knowledge or expertise. However, the more you have built your sense of self around a particular form of professional expertise, the more threatened you are by a technology that challenges that expertise. And the less equipped you may be to reconstruct your identity around something different, because you have never had to.

The junior or unskilled employee, paradoxically, is often better placed to adapt. They are less invested in the existing hierarchy. Less defined by a specific professional role. More accustomed to not yet being the expert. More open, because they have less to lose, to a world in which the definition of expertise changes. And for the corporate grifters, AI becomes a cover for their lacking expertise.

This is the twist that the AI conversation in organisations consistently underestimates. The resistance that skilled, experienced people offer is not a change management problem to be overcome. It is an identity crisis to be understood.

What Is Being Asked

When an organisation implements AI and asks its experienced professionals to adopt it, on the surface it is asking them to learn a new tool, integrate it into their workflow and leverage its capabilities to work more effectively. Underneath the surface it is asking them to redefine what it means to be good at their job. To reconstruct their sense of professional identity around a new relationship between human judgment and machine capability. To accept that the expertise they have spent decades building, which has been the primary source of their professional status and self-worth, is no longer the scarce and valued resource it once was.

The people being asked to do it are being asked to do it in organisations that have rarely, if ever, provided the support, the space or the language to navigate that kind of transition. Worse yet, the change management frameworks that these organisations are using were only ever designed to address the surface request. They provide training, they communicate benefits, they manage stakeholder concerns and they track adoption metrics.

They are entirely inadequate to support an existential transformation.

What A Human Response Looks Like

So how should we be handling this?

A human response to AI adoption resistance begins by acknowledging what is actually being resisted. Offering an acknowledgement at the start of the process changes the nature of the conversation and how it is conducted.

It means that the first conversation about AI adoption in a team or organisation should not be about the tools, the timeline or the productivity targets. It should be about what people value in their work, what gives them a sense of pride and accomplishment, what they are most afraid of losing, what they hope the technology might free them from and what they hope it will never be allowed to replace.

That level of empathy and engagement would pave the way for adoption that produces genuine value rather than surface compliance simply because when people understand what is changing about their work and have had a genuine voice in shaping how it changes, they will actively participate in the journey rather than be dragged along by it.

It means investing in what researchers call human-AI augmentation, being the design of workflows that enhance rather than replace human judgment, that make the experienced professional more capable rather than more redundant, that place the expertise accumulated over decades in a relationship with AI capability that amplifies rather than devalues it.

It means treating agency as a design principle. Not involving people in implementation after the decisions have been made, but involving them in the decisions themselves. Asking not how do we get people to use this but what would this need to look like for you to want to use it.

And it means, perhaps most importantly, acknowledging openly that the transition is genuinely difficult. That the experienced professional who feels their expertise is being devalued is not wrong to feel that. That their sense of loss is real and legitimate and deserves more than a change management programme designed to overcome it.

The Question Underneath The Resistance

It is worth considering that when a professional pushes back, resists or disengages, they are really asking is:

It is a powerful question that deserves a considerate, truthful answer that acknowledges what they have built; the judgement, the contextual knowledge, the capacity to understand what a client or patient or colleague actually needs in a way that no aggregate model can replicate and that is increasingly rare and valuable.

It is the human layer that gives AI capability meaning and direction.

The experienced professional who knows when the AI is wrong. Who understands the context that the model cannot see. Who can make the judgment call that the output is technically correct but humanly inadequate. That person is not redundant. They are essential. But they need to be enabled to occupy that role rather than having it designed away from them in the rush to implement.

The organisations that will navigate the AI transition most effectively will be the ones that take that resistance seriously enough to understand what it is really about.

It is a battle over identity.