What is reactance in AI governance for community organisations?
Reactance AI governance describes the motivational resistance that staff experience when AI systems or governance frameworks are perceived as threatening their professional autonomy. Community organisations and not-for-profits need to understand reactance to design AI governance that builds engagement rather than triggering resistance.
Reactance
Reactance AI governance addresses one of the most common and least acknowledged reasons that AI governance fails in practice. When staff experience AI tools or governance requirements as threats to their professional autonomy and judgement, they do not simply comply less enthusiastically. Rather, they resist. Reactance is the motivated push back against perceived restrictions on freedom — and in community organisations, where professional identity is closely tied to autonomous judgement and relational practice, it is a predictable response to AI implementation that does not account for it.
How does psychological reactance affect not-for-profit AI governance and staff adoption?
Definition
What is Reactance in AI Governance?
Psychological reactance describes the motivational state that arises when a person perceives that their freedom of behaviour is being threatened or eliminated. When people experience reactance, they become motivated to restore the threatened freedom — and the more important the freedom, the stronger the motivation to restore it. Crucially, reactance does not merely reduce compliance. It actively reverses it. People experiencing reactance are more likely to do the opposite of what is being asked, precisely because doing so restores their sense of autonomous agency.
In AI governance contexts, reactance is triggered when staff perceive that AI tools or governance requirements are restricting their professional judgement rather than supporting it. Furthermore, this perception does not need to be accurate. Reactance is a response to the experience of constraint, not to the objective reality of it. Consequently, governance frameworks that are genuinely designed to support professional practice can still generate reactance if they are experienced as controlling, if implementation is top-down, or if staff have not been meaningfully involved in designing how AI tools are used in their work.
The Boomerang Effect
Research consistently shows that reactance can produce a boomerang effect: persuasive attempts that are perceived as controlling or freedom-threatening produce the opposite of their intended result. In AI governance contexts, this means that mandates, requirements, and compliance-oriented implementation approaches are more likely to generate resistance than voluntary or autonomy-supportive ones. Moreover, the boomerang effect is stronger when the threatened freedom is closely tied to professional identity, which is precisely the case for community sector workers whose sense of professional competence rests on autonomous relational judgement.
How this differs from simple resistance: General resistance to change is a broad category that includes inertia, unfamiliarity, and lack of confidence. Reactance is a specific motivational response to perceived threats to autonomy. The governance implication is different in each case: resistance from inertia is addressed by support, training, and time. Reactance, by contrast, is addressed by restoring or preserving the sense of autonomous choice — which means that adding more support, training, and encouragement can make it worse if those additions are also experienced as controlling.
Research Origin
Brehm (1966) established the foundational theory of psychological reactance, proposing that people experience a motivational state of reactance when their freedoms are threatened or eliminated, and that this state generates specific and predictable responses aimed at restoring those freedoms. In addition to engaging in the restricted behaviour, reactance responses include increased subjective valuation of the threatened freedom, negative affect directed toward the source of the threat, and, in cases where direct restoration is impossible, engagement in behaviourally similar alternatives. Applied to AI governance, Brehm’s framework predicts that staff who experience AI implementation as threatening their professional autonomy will be motivated to resist, circumvent, or devalue AI tools and governance requirements — not from laziness or obstruction, but from the fundamental human motivation to maintain freedom of action.
Miron and Brehm (2006) reviewed four decades of reactance research, confirming the robustness of the effect across contexts and identifying several important moderating conditions. Their analysis showed that reactance is intensified when the threatened freedom is important to the person, when the threat is perceived as illegitimate or unjustified, and when the person has a strong sense of personal agency. In community sector AI governance contexts, all three conditions are often present: professional autonomy is central to practitioner identity, AI implementation is frequently top-down rather than co-designed, and community workers typically have a strong sense of professional agency developed through training and practice.
Siegel and Rosenberg (2025) provide a current introduction and overview of reactance theory, noting that the field has established practical implications for message design and intervention: autonomy-supportive language, choice-enhancing framing, and approaches that acknowledge the person’s right to decide are consistently more effective at achieving desired outcomes than directive or controlling ones. These findings have direct implications for how AI governance requirements are communicated and implemented in community organisations.
Further reading: The Decision Lab provides an accessible overview of reactance theory and its applications across organisational and health communication contexts. The Decision Lab: Reactance Theory ↗
In AI Contexts
How Reactance AI Governance Risks Operate
Reactance AI governance risks develop through two distinct pathways. In the first, AI tools are perceived as replacing or overriding professional judgement, generating reactance against the tools themselves. Staff may avoid using AI tools, find reasons to reject their outputs, or develop workarounds that bypass governance requirements. In the second pathway, governance frameworks themselves are perceived as controlling, generating reactance against the governance process. Staff may comply minimally, find ways to satisfy requirements without genuinely engaging with them, or become less willing to raise concerns about AI use because the governance process feels like surveillance rather than support.
Mandatory Compliance and Reactance Escalation
Mandatory AI use policies are a particularly strong trigger for reactance in professional contexts. When staff are required to use AI tools rather than invited to do so, the requirement itself signals a restriction on the professional freedom that trained practitioners typically hold over the methods they use in their practice. Consequently, mandatory adoption can generate reactance even among staff who would otherwise be open to or positive about AI tools, because the manner of implementation has converted a potentially autonomous choice into an externally imposed constraint.
Furthermore, reactance can escalate when governance responses to initial resistance become more controlling. An organisation that responds to staff reluctance with stronger mandates, monitoring, and compliance requirements may find that resistance intensifies rather than diminishes, precisely because each escalation is experienced as a further threat to autonomy and generates additional reactance motivation.
Autonomy-Supportive Implementation as a Governance Design Principle
Research on reactance points consistently toward autonomy-supportive approaches as more effective than directive ones. In practice, this means framing AI tools as expanding rather than constraining professional options, involving staff in selecting and configuring tools, and designing governance requirements in ways that acknowledge and preserve professional discretion. Moreover, providing genuine choices — including the choice not to use a particular tool in a particular context — significantly reduces reactance even when overall governance expectations are maintained.
A practical governance test: Review how AI governance requirements are communicated to staff. Count the number of times the language is directive (“must”, “required”, “mandatory”) versus autonomy-supportive (“you might consider”, “in most cases”, “where appropriate”). Research on reactance shows that language alone significantly affects the motivational response it generates. Governance requirements can be equally clear and firm in both registers — but the autonomy-supportive register is substantially less likely to trigger the resistance that undermines implementation.
Why It Matters for Not-For-Profits
The Specific Risk for Community Organisations
Community sector workers typically have strong professional identities built around autonomous relational practice. Social workers, community health practitioners, peer support workers, and allied health professionals have spent years developing professional judgement that is not rule-based but contextual, relational, and responsive to individual client circumstances. Consequently, AI tools that appear to reduce that contextual judgement to algorithmic outputs, or governance frameworks that appear to subject professional practice to algorithmic oversight, are particularly likely to trigger reactance in these workforces.
Peer Workers and Reactance to Algorithmic Authority
For peer workers and lived experience staff, reactance to AI governance may be particularly pronounced. Their practice authority rests explicitly on personal experience and relational authenticity rather than professional credentials or procedural compliance. In contrast, AI systems derive authority from data and algorithms. These are fundamentally different bases for practice authority, and staff whose practice rests on the first may experience AI-derived outputs as a challenge not just to their professional freedom but to the legitimacy of their practice model itself. Accordingly, governance responses that acknowledge this tension rather than dismissing it are substantially more likely to succeed.
Co-Design as a Structural Reactance Reduction Strategy
The most robust structural response to reactance in AI governance is co-design: involving staff meaningfully in decisions about which AI tools are used, how they are integrated into practice, and what governance requirements apply. Co-design reduces reactance not by making requirements less stringent but by converting externally imposed constraints into self-generated ones. Research consistently shows that people experience far less reactance to restrictions they have participated in creating, even when those restrictions are substantively identical to ones they would resist if imposed from outside.
A starting point: Before implementing any new AI governance requirement, identify two or three staff members who are likely to be most affected and ask them directly: “Does this feel like it respects your professional judgement, or does it feel like it replaces it?” Their answer will tell you whether the requirement is likely to generate reactance, and give you the information you need to adjust the design or the framing before implementation rather than after resistance has already developed.
Reactance AI governance describes the motivational resistance that staff experience when AI systems or governance frameworks are perceived as threatening their professional autonomy. Community organisations and not-for-profits need to understand reactance to design AI governance that builds engagement rather than triggering resistance.
Observable Signals
What to Watch For
These patterns suggest reactance may be shaping staff responses to AI tools or governance in your organisation. They are signals worth naming and discussing, not a definitive checklist.
- Staff who were generally open to change and new tools are actively resistant to AI implementation in ways that seem disproportionate to their usual approach, suggesting that the manner of introduction — top-down, mandatory, or perceived as replacing professional judgement — has triggered reactance that would not have developed under a different implementation approach.
- Staff are finding technically compliant ways to satisfy AI governance requirements without genuinely engaging with them — completing required documentation without substantive review, ticking boxes without applying the judgement those boxes are designed to capture — indicating that governance requirements are being experienced as impositions to be managed rather than as practices to be genuinely adopted.
- Organisational responses to initial staff reluctance about AI tools have involved stronger mandates, more monitoring, or additional compliance requirements, and resistance has increased rather than decreased in response, which is the characteristic escalation pattern of reactance when governance approaches become progressively more controlling.
- Staff describe AI tools or governance requirements in terms of professional autonomy — “it takes away my judgement”, “it doesn’t trust us to make decisions” — rather than in terms of practical usability or effectiveness, indicating that the reactance trigger is the perceived threat to professional freedom rather than specific tool features or workload concerns.
Reactance AI governance and professional autonomy in community sector AI implementation
Reactance AI governance describes the motivational resistance that staff experience when AI systems or governance frameworks are perceived as threatening their professional autonomy. Community organisations and not-for-profits need to understand reactance to design AI governance that builds engagement rather than triggering resistance.
External References
- Brehm, J.W. (1966). A theory of psychological reactance. Academic Press.
- Miron, A.M. and Brehm, J.W. (2006). Reactance theory — 40 years later. Zeitschrift fur Sozialpsychologie, 37(1), 9-18.
- Siegel, J.T. and Rosenberg, B.D. (2025). Psychological reactance theory: An introduction and overview. Motivation Science.
- Steindl, C., Jonas, E., Sittenthaler, S., Traut-Mattausch, E. and Greenberg, J. (2015). Reactance theory. In Encyclopedia of Personality and Individual Differences. Springer.
- The Decision Lab. Reactance theory: Reference guide.
Related Lexicon Entries
From theory to practice
Understanding the risk is the first step.
The Not For Humans Toolkit gives your organisation practical tools to design AI governance implementation that supports professional autonomy rather than threatening it, build co-design processes that convert governance requirements from external impositions into self-generated commitments, and develop communication approaches that frame AI tools in autonomy-supportive language that reduces rather than triggers reactance. The Behaviour Risk Cards name reactance in language your whole team can use.
Not For Humans: Because AI’s greatest risk isn’t technical, it’s Human.