The Motivation Problem and Getting People to Truly Give AI a go!
Seligman’s Learned Helplessness and Deci & Ryan’s Self-Determination Theory
Every enterprise that has attempted more than one transformation programme carries invisible scar tissue. It is not in the strategy documents or the retrospectives. It is in the people; in what they have learned, through repeated experience, about the relationship between their effort and any outcome that matters.
Martin Seligman, an American psychologist who founded the positive psychology movement, spent decades studying what happens when people learn that their actions have no effect. Edward Deci and Richard Ryan, working in parallel at the University of Rochester, spent equally long studying what happens when people’s fundamental psychological needs are met, or crushed. Read separately, each is illuminating. Read together, they provide a useful diagnosis of why transformation programmes generate compliance instead of commitment, and why that difference determines everything.
1. Learned Helplessness: The Rational Decision to Stop Trying
Seligman’s foundational insight emerged from experiments that, by modern standards, were brutal in their simplicity. Animals exposed to inescapable shock learned to stop attempting escape, even when the conditions changed and escape became possible. The mechanism was not emotional but cognitive: the subjects had learned a mental model. My actions do not affect outcomes. Once this model was established, it generalised. Helplessness learned in one context bled into unrelated domains, producing passivity, cognitive deficits, and what Seligman initially described as analogous to depression.
The transfer to organisational life is sobering.
Employees who have repeatedly proposed improvements that were ignored learn to stop proposing.
Teams that have survived three “transformational” programmes, each launched with fanfare, each quietly abandoned, learn that the safest strategy is to wait for the enthusiasm to pass while appearing to comply. Middle managers who have watched their honest assessments of delivery risk be overruled by executive optimism learn to produce the numbers executives want to see. We’ve seen it all before and as leaders, we have perhaps caused it too.
This is not cynicism. It is learning. And it is precisely the kind of learning that Argyris described as the enemy of organisational adaptation: a single-loop response that protects the individual but prevents the organisation from confronting its own dysfunction. The defensive routines Argyris documented - the skilled ways that smart people avoid having their reasoning examined - are, in Seligman’s terms, the coping strategies of the organisationally helpless. They have learned that questioning the programme is futile, so they learn instead to navigate the programme without exposing themselves to risk.
For AI transformation, the implication is specific. When you announce the AI strategy and notice that the response from experienced staff is polite attentiveness followed by minimal engagement, you are not observing resistance. You are observing learned helplessness and the learning was rational. These people sat through the Agile transformation. They sat through the DevOps transformation. (I lead many DevOps transformations, so this really hurts to admit!) They noticed that each programme promised to change the way the organisation works, and each programme was absorbed by the existing structures until what remained was the terminology without the transformation. Ralph Stacey would say: the gesture was made, and the response of the organisation was to neutralise it. The people closest to the work noticed. And they drew the correct conclusion (for them).
2. Explanatory Style: The Story People Tell Themselves About Why Things Fail
Seligman’s later work moved from helplessness to its opposite, resilience. The bridge between them is what he called explanatory style. This is the habitual way people explain bad events to themselves, and it operates along three dimensions.
Permanence: “This will never change” versus “This is a setback.”
Pervasiveness: “Everything is broken” versus “This particular thing went wrong.”
Personalisation: “It is my fault” versus “The circumstances were against us.”
A pessimistic explanatory style which is permanent, pervasive, personal, indicates helplessness. An optimistic style which is temporary, specific, external for bad events, predicts persistence and recovery.
The organisational application connects directly to Dweck’s mindset research. Dweck showed that beliefs about ability determine responses to failure. Seligman adds the mechanism by which those beliefs are constructed: through the stories people tell themselves about why things went wrong. In a fixed mindset culture where the explanatory style is pessimistic, a failed pilot is not “an experiment that revealed something about our requirements” (temporary, specific, external). It is “proof that we are not capable of this” (permanent, pervasive, personal). The fixed mindset provides the belief system; the pessimistic explanatory style provides the narrative engine that reinforces it after every setback.
Karl Weick’s sensemaking theory illuminates why explanatory style matters so much in organisations. Sensemaking is retrospective - people look back at what happened and construct a plausible narrative. The explanatory style is the filter through which that retrospective narrative is constructed. Two teams can have the same experience and construct entirely different meanings from it. One team says: “Our specification was ambiguous in section three; we need to tighten the constraints.” The other says: “AI-generated code cannot be trusted; we need more governance.” Same event. Different sensemaking. Different future.
The leader’s task, in Seligman’s terms, is not to impose optimism.
Forced positivity in the face of genuine dysfunction is gaslighting, not leadership.
The task is to model a specific explanatory discipline: when things go wrong, be temporary (”this iteration did not work”), specific (”the input schema was underspecified”), and external to identity (”the specification needs revision, not the team”). This is not spin. It is the practice of accurate attribution and it is far harder than it sounds, because in most organisations the default explanatory style for failure is the opposite: permanent (”we always struggle with this”), pervasive (”the whole programme is in trouble”), and personal (”someone needs to be held accountable”).
Dekker’s Just Culture framework operationalises this shift. His substitution test - “Would another similarly trained person in the same circumstances have done the same thing?” - is an explanatory style intervention. It moves attribution from personal to systemic, from permanent character flaw to temporary situational constraint. It does not remove accountability. It redirects accountability from individual blame to systemic learning. Seligman would recognise this as the organisational equivalent of cognitive restructuring, changing the explanatory habit that sustains helplessness.
3. Self-Determination Theory: The Three Needs That Transformation Programmes Systematically Destroy
Deci and Ryan’s Self-Determination Theory identifies three innate psychological needs whose satisfaction produces intrinsic motivation and whose frustration produces disengagement, regardless of external incentives.
Autonomy is the need to feel that one’s actions are self-directed and freely chosen. This is not necessarily independence, a person can autonomously choose to follow instructions if they understand and endorse the reasons. Autonomy is about volition: the experience of choosing rather than being controlled.
Competence is the need to feel effective in one’s interactions with the environment. It is the desire for challenge at the right level; not so easy that it bores, not so hard that it overwhelms. Positive feedback that is informational supports competence. Negative feedback that is judgmental destroys it.
Relatedness is the need to feel connected to and cared about by others. It does not require deep friendship, mutual respect and shared purpose are sufficient. But isolation and exclusion are very demotivating regardless of other conditions.
The synthesis with Seligman is this: learned helplessness is what happens when all three needs are thwarted simultaneously over time. When your actions have no effect (autonomy destroyed), when your existing skills become irrelevant and no support is provided for new ones (competence threatened), and when the team structures you relied on are disrupted without replacement (relatedness severed) the rational response is disengagement. Seligman describes the outcome. Deci and Ryan describe the mechanism.
Large scale tech transformation (Agile, DevOps, AI, whatever) as typically implemented in large enterprises, thwarts all three needs with remarkable efficiency.
Autonomy is destroyed when governance frameworks dictate precisely how AI may be used, when tool selection is centralised, when experimentation requires approval through three layers of management, and when the transformation itself is mandated rather than invited. Tom Peters would recognise this immediately: the bureaucratic reflex to control the new technology before anyone has understood what it can do. The irony is that the governance designed to manage risk is itself the primary risk because it extinguishes the intrinsic motivation that is the only reliable fuel for learning.
Competence is threatened in two directions simultaneously. Existing competence, the skills that took years to develop and that provide daily experiences of mastery, that form the basis of professional identity, is devalued. A senior developer who experiences flow while writing elegant code is told that their future is writing specifications. This is not a neutral skill transition. It is what Anthony Giddens calls a disruption of ontological security: the stable sense of self that comes from routines that confirm who you are. At the same time, the new competence is not yet available. The developer does not yet know how to write precise specifications. They are, temporarily, incompetent — and in a fixed mindset culture, temporary incompetence is experienced as permanent inadequacy.
Relatedness is disrupted when team structures change, when new roles create unfamiliar reporting lines, when the people you learned to trust are reassigned, and when the transformation creates a perceived in-group of “AI enthusiasts” and an out-group of everyone else. Stacey would say that the pattern of interaction that constituted the team’s identity has been broken, and no new pattern has yet formed. In the gap, people feel disconnected and disconnection compounds the effects of lost autonomy and threatened competence.
4. The Motivation Continuum: Why “Mandated Change” Is an Oxymoron
Deci and Ryan’s most practically useful contribution is their description of motivation as a continuum, not a binary. Between pure extrinsic motivation (doing something because you are told to and will be punished if you do not) and pure intrinsic motivation (doing something because it is inherently satisfying) lie several intermediate stages:
External regulation: “I will use the AI tool because my manager told me to.” This is compliance, and it disappears the moment the oversight does.
Introjected regulation: “I should use the AI tool because I will feel guilty if I do not.” This is internalised pressure, and it produces anxiety rather than engagement.
Identified regulation: “I use the AI tool because I have thought about it and believe it aligns with what I value in my work.” This is genuine buy-in where the person has made the goal their own, even if the activity is not intrinsically enjoyable.
Integrated regulation: “Using AI tools is consistent with who I am as a professional and what I believe about the future of my craft.” This is full internalisation and the change has become part of identity.
Drucker’s insight about knowledge workers is the organisational expression of this continuum. Knowledge workers must define the task before they can do it. They own the means of production and carry it with them when they leave. They cannot be supervised in the Taylorist sense because the supervisor cannot evaluate work they do not understand (as fully as they do). Autonomy, in Drucker’s terms, is not a motivational perk, it is a structural requirement of knowledge work. Deci and Ryan provide the psychological explanation for why Drucker’s observation is true:
knowledge work requires intrinsic motivation, and intrinsic motivation requires autonomy, competence, and relatedness.
5. The Synthesis: Reversing Helplessness by Satisfying Needs
Seligman and Deci & Ryan, read together, provide a diagnostic and a prescription.
The diagnostic: in any organisation with a history of failed transformation programmes, you are dealing with a population whose learned helplessness is rational, whose psychological needs are being systematically thwarted, and whose apparent “resistance” is actually the most adaptive response available to them given the conditions they face.
The prescription: reverse the helplessness by creating conditions that satisfy the three basic needs.
Restore autonomy by offering genuine choice. Not the false choice of “adopt the mandated tool or explain why not,” but the real choice of which problems to tackle, which approaches to try, and how to evaluate results. Mintzberg’s emergent strategy provides the framework: instead of designing the AI adoption plan and instructing teams to execute it, create the conditions for teams to experiment and detect the patterns that emerge. The actual strategy will be the pattern across those experiments.
Rebuild competence through mastery experience. The most powerful source of self-efficacy is actually doing the thing and succeeding. Not being told you can do it. Not watching a vendor demonstration. Doing it yourself. The specification-generate-validate loop is a competence engine when it works. The specification provides a clear goal, the generated output provides immediate feedback, and increasingly complex specifications provide progressive challenge. These are precisely Csikszentmihalyi’s conditions for flow. But this potential is only realised if the first experience is achievable. Seligman’s research shows that even a small number of experiences where effort produces results can reverse learned helplessness but the experiences must be authentic, not manufactured.
This is why Weick’s small wins are not merely a tactical recommendation but a motivational necessity. Each small win is simultaneously a sensemaking event (Weick), a competence-satisfying moment (Deci and Ryan), and an evidence point against the learned helplessness narrative (Seligman). A team that writes a specification and watches AI generate working code from it has experienced something that no strategy document can provide: the felt connection between their effort and an outcome. That felt connection is the antidote to helplessness.
Protect relatedness by making AI adoption a team practice, not an individual mandate. Pair specification writing. Team-based validation. Shared learning sessions where people reflect together on what worked and what did not. Communities of practice that connect people across teams who are navigating the same learning curve. The goal is not merely knowledge sharing it is the maintenance of the social bonds that sustain motivation when competence is temporarily disrupted and autonomy is under pressure.
6. The Extrinsic Reward Trap: When Incentives Backfire
Deci and Ryan’s research contains a finding that most transformation leaders find counterintuitive: extrinsic rewards can undermine intrinsic motivation. This is the overjustification effect, when you pay someone to do what they previously did for interest, they reframe the activity as work. Remove the reward and the activity stops. The reward has become the reason.
This does not mean you should not pay people. It means that contingent rewards for activities that could be intrinsically motivating are dangerous. Offering bonuses for “AI adoption targets” is a possible example: it reframes engagement with AI from “an interesting new capability worth exploring” to “something management wants me to do for which I will be compensated.” The extrinsic reward displaces whatever intrinsic interest might have existed.
Peters intuited this decades before the experimental evidence was clear. His insistence that excellence requires passion, that procedure kills intrinsic motivation and replaces it with compliance. It is the managerial expression of what Deci and Ryan demonstrated in the laboratory - Bureaucracy does not merely slow organisations down, it systematically converts intrinsic motivation into external regulation, and in doing so, destroys the only form of motivation that produces genuine learning.
The alternative is what Deci and Ryan call autonomy-supportive leadership: explaining the rationale for the change rather than simply mandating it; acknowledging that the transition is difficult, providing choice within structure and offering informational feedback rather than controlling evaluation. This is not soft management. It is the only approach that produces the identified and integrated regulation that transformation requires.
7. The AI-Specific Challenge: When the Tool Itself Threatens the Need
There is a dimension of AI adoption that makes the motivation problem uniquely acute. AI does not merely change how people work. It changes the basis of their competence. A developer whose mastery experience came from writing elegant code faces a tool that writes code. A data analyst whose professional identity was built on the ability to extract insight from complex datasets faces a tool that does the same. The threat is not to their job in the immediate term, it is to their competence need, which is more psychologically fundamental than their employment status.
The response cannot be “reskilling” as typically practised; a training course that provides information without providing mastery experience, followed by an expectation that people will immediately be competent in the new approach. I am sure we have all seen this with disastrous “agile” transformations. Seligman’s research on helplessness reversal shows that the transition must be graduated, that early successes must be genuine, and that the connection between effort and outcome must be visible and immediate.
Dweck’s growth mindset is the cultural precondition: people must believe that the new competence is acquirable through effort. Deci and Ryan provide the environmental conditions: autonomy in how they acquire it, optimal challenge that stretches without breaking, and relatedness that means they are not alone in the learning. Seligman provides the recovery mechanism: small, authentic experiences of mastery that rebuild the belief that effort produces results.
Together, these three frameworks describe a transformation approach that looks very different from the standard model. Instead of a mandated programme with adoption targets and compliance dashboards, it looks like this:
Teams are invited (not instructed) to experiment with AI for problems they choose (not problems assigned to them).
The first experiments are designed to succeed - not artificially, but by selecting problems where the specification-generate-validate loop is likely to produce visible results quickly.
The learning is social - people work in pairs, reflect in teams, and share across communities. The explanatory style when things go wrong is specific, temporary, and systemic and the extrinsic incentives are informational.
(An Organisational Prompt is something you can do now…)
Organisational Prompt
Ask a team you’re wanting to get going on AI a question: “If you could use AI to solve one problem that actually matters to you, not the organisation, what would it be?”
This is Deci and Ryan’s intervention: you are restoring autonomy (their choice), targeting competence (a problem they understand deeply), and building relatedness (you are listening, which is itself a relational act).
Further Reading
Martin Seligman: Learned Optimism: How to Change Your Mind and Your Life - The accessible version of Seligman’s research on helplessness and its reversal. Read it for the explanatory style framework, which is the most practically useful tool for leaders who need to diagnose why their people have stopped engaging.
Edward Deci and Richard Ryan: Self-Determination Theory: Basic Psychological Needs in Motivation, Development, and Wellness - The comprehensive academic statement of SDT. Dense but essential.
Daniel Pink: Drive: The Surprising Truth About What Motivates Us - The popular synthesis of Deci and Ryan’s work applied to the workplace. Useful as an introduction, but read the original research for the nuance that Pink necessarily simplifies.
Albert Bandura: Self-Efficacy: The Exercise of Control - The research on why mastery experience is the most powerful source of belief in one’s own capabilities. The bridge between Seligman’s helplessness reversal and Deci and Ryan’s competence need.
Disclaimer
I write about the industry and its approach in general. None of the opinions or examples in my articles necessarily relate to present or past employers. I draw on conversations with many practitioners and all views are my own.
Point 7 is more salient, as AI or process reengineering is generally implemented with an implicit cost-cutting mandate rather than a productivity mandate. It needs to be clear to staff that implementation is not about replacement, but about enabling the completion of the large backlog of demand.