The same AI rollout can lead one employee to rethink how work gets done while leaving another anxious, disengaged and increasingly resistant to change. The difference often lies less in the technology itself than in how employees interpret what the technology means for their own work, and those interpretations can shift surprisingly quickly.
The Leadership Question Organizations Are Often Asking the Wrong Way
One of the more puzzling aspects of AI transformation is the variation organizations observe after what appears to have been a successful rollout. Similar teams receive access to the same technology, participate in the same training, and operate under the same strategic objectives, yet their responses diverge in ways that are difficult to explain. Some employees begin experimenting almost immediately, gradually redesigning how they work and identifying opportunities that had not previously existed. Others become noticeably more cautious. Adoption slows, enthusiasm fades and managers often conclude that they are facing resistance to change.
This explanation is understandable, but it may also be incomplete. It assumes that employees hold relatively stable attitudes towards AI and that the task of leadership is largely to distinguish between those who embrace the technology and those who remain sceptical. In practice, our experience suggests that people’s responses are considerably more dynamic than that. What determines whether AI becomes a source of innovation or a source of anxiety is often not the technology itself, but the meaning employees attach to it within the context of their everyday work.
A recent diary study published in Group & Organization Management reinforces this observation in a particularly interesting way.
Why the Same Person Can Experience AI Very Differently
Rather than comparing different groups of employees, the researchers followed the same individuals across consecutive working days. The resulting picture is considerably more nuanced than many organizations assume. The distinction did not emerge between employees who consistently embraced AI and those who consistently rejected it. Instead, the same individual could experience AI as a challenge on one day and as a threat on another, depending on workload, task complexity and their broader psychological context.
This finding deserves attention because it shifts the leadership question entirely. If employees’ responses fluctuate over time, resistance can no longer be understood simply as a personality characteristic or a cultural trait. It becomes a contextual phenomenon. The more useful question therefore becomes: under what conditions are employees likely to experience AI as an opportunity to expand their contribution, and under what conditions does the same technology begin to feel like a threat to their expertise, identity or future role?
That distinction is considerably more actionable because organizational conditions are, at least in part, shaped by leadership.
Why the Distinction Between Challenge and Threat Matters
The study is valuable not simply because it identifies two different attitudes towards AI, but because those attitudes are associated with fundamentally different patterns of behaviour.
Employees who experienced AI as a challenge were significantly more likely to engage in what researchers describe as task crafting. Rather than using AI simply to complete familiar tasks more efficiently, they began reconsidering how those tasks could be organised, redirecting newly available cognitive capacity towards higher-value work, and experimenting with different ways of creating value for the organization.
Employees who experienced AI primarily as a threat followed a markedly different trajectory. Job replacement anxiety became more pronounced, cynicism increased and engagement declined over time. Perhaps the most interesting finding, however, was that reducing anxiety alone did not automatically produce innovation. Positive attitudes towards AI appeared to buffer employees against the threat response, but innovative behaviour emerged only when people also had sufficient freedom to act on the opportunities the technology created.
This distinction is particularly relevant because many organizations continue to invest heavily in communication and training while paying comparatively little attention to the conditions that determine whether employees can meaningfully redesign their work once AI creates additional capacity.
The Missing Piece Is Often Work Design
This is where the research aligns closely with what we observe in organizations undergoing AI transformation.
Employees rarely respond to AI as a neutral piece of software. They respond to what they believe the technology implies about their future contribution, their professional identity and the degree of influence they retain over their own work. Those interpretations are continuously shaped by managerial behaviour, organizational communication, workload and everyday experience.
For that reason, autonomy deserves considerably more attention than it typically receives during AI implementation. An employee may understand the technology perfectly well and genuinely appreciate its potential, yet still have little opportunity to use it creatively if workflows remain tightly prescribed and success continues to be measured exclusively through compliance with existing processes. Under those conditions, AI accelerates established routines but leaves little room for experimentation. Where employees are trusted to rethink aspects of their own work, the additional cognitive capacity created by AI is far more likely to be invested in better decisions, improved collaboration and more innovative ways of solving problems.
The difference is subtle but important. AI creates additional capacity. Leadership determines whether that capacity is simply absorbed by existing workflows or translated into new forms of value.
What This Means for Leaders
One implication of the study deserves particular attention because it changes how managers should think about their role during AI transformation.
Organizations often assume that once employees have learned how to use AI, adoption becomes largely a matter of individual choice. The evidence suggests something considerably more dynamic. Employees continuously reassess what AI means for their work as organizational conditions evolve around them. A demanding week, an unclear conversation about changing expectations, a manager who rewards efficiency but discourages experimentation, or a workflow that leaves little room for independent judgement can all shift that interpretation from challenge to threat.
The manager’s role therefore extends well beyond implementation. Through everyday decisions about workload, autonomy, feedback and expectations, managers shape the psychological conditions under which employees encounter AI. Those conditions, in turn, influence whether employees use the technology merely to execute familiar tasks more quickly or to rethink how work itself could be performed.
Looking Beyond Adoption Metrics
Most organizations devote significant effort to measuring AI adoption through indicators such as training completion, system usage or productivity improvements. These measures are undoubtedly useful, but they reveal relatively little about whether AI is changing behaviour in ways that create lasting organizational value.
In our work with organizations, we encourage leaders to look beyond adoption and ask a different set of questions. Are employees proactively redesigning how work gets done? Do managers create sufficient psychological safety for experimentation? Where is additional cognitive capacity being converted into better decisions rather than simply more activity? Which teams consistently transform AI into innovation, and which continue using it only to accelerate existing routines?
These questions provide a far richer picture of whether AI is becoming embedded in the way an organization thinks and works, rather than simply becoming another tool employees have learned to use.
The Leadership Opportunity
Perhaps the most encouraging implication of the research is that employees’ responses to AI are not fixed. They evolve as organizational conditions evolve. Workload changes. Managerial behaviour changes. People’s confidence grows or diminishes through everyday experience. The interpretation employees attach to AI develops alongside those experiences.
For leaders, this changes the nature of the challenge. Rather than trying to convince employees that AI is either good or bad, the more valuable task is to create an environment in which people are able to experience the technology as an opportunity to contribute more effectively, exercise greater judgement and engage in more meaningful work.
Organizations often describe AI transformation as a technology programme supported by change management. Our experience suggests that the opposite perspective is often more useful. AI transformation is fundamentally a leadership challenge that happens to involve technology. The organizations creating the greatest value are rarely those with access to fundamentally different tools. More often, they are those that understand how leadership, work design and human behaviour determine whether technological capability ultimately becomes organizational performance.