Trust Calibration
Research-grounded · Original operationalization · Construct validation planned · v0.3
RIST Trust Diagnostic
Four dimensions of trust that predict successful AI adoption.
Four dimensions of trust that predict successful AI adoption — all require calibration, not maximization. Trust exists on a spectrum: undertrust and overtrust are both failure modes. Rate each question on a 1–5 frequency scale. All questions are behavioral: they ask what you do or observe, not what you think or believe.
1
Relational Trust
Trust between humans in the adoption process. Will AI threaten my livelihood, status, or relationships? Both undertrust (resistance, sabotage) and overtrust (blind compliance) are failure modes.
Undertrust Question
Q1How often do you express concerns about AI directly to a colleague or manager rather than discussing them privately with others?
Undertrust Question
Q2When a colleague shares concerns about AI, how often do you ask what might address those concerns rather than dismissing them?
Overtrust Question
Q3How often do you adopt a new AI tool without understanding how it affects your team's work or relationships?
Overtrust Question
Q4How often do you implement an AI solution because "leadership says we should" without considering impact on team trust?
2
Institutional Trust
Trust in the organization's governance, technology infrastructure, and policies around AI. Is the organization doing this responsibly? Both undertrust (shadow AI, workarounds) and overtrust (naive faith in policy docs) are failure modes.
Undertrust Question
Q1How often do you question whether the organization's AI governance is actually protecting employee interests?
Undertrust Question
Q2When the organization introduces an AI policy, how often do you assume it's incomplete or will be bypassed?
Overtrust Question
Q3How often do you trust organizational AI policy without independently verifying what the system actually does?
Overtrust Question
Q4How frequently do you assume compliance and security are handled correctly based on official documentation alone?
3
Self-Trust
Trust in your own ability to learn, adapt, and remain relevant in an AI-augmented environment. Can I learn this? Am I becoming obsolete? Both undertrust (learned helplessness, avoidance) and overtrust (overestimating fluency) are failure modes.
Undertrust Question
Q1How often do you avoid learning AI tools because you believe you're "not a technical person"?
Undertrust Question
Q2When you struggle with a new AI tool, how often do you assume you're not capable rather than trying a different approach?
Overtrust Question
Q3How often do you skip training or documentation because you assume you already understand how AI works?
Overtrust Question
Q4How frequently do you trust your ability to use an AI tool without checking whether you're using it correctly?
4
Task Trust
Trust in the AI output for a specific task. Is this output reliable? Can I stake my reputation on it? Both undertrust (checking everything manually) and overtrust (accepting outputs uncritically) are failure modes.
Undertrust Question
Q1How often do you verify AI-generated outputs by redoing the work manually rather than using the output?
Undertrust Question
Q2When an AI tool provides an answer, how frequently do you distrust it enough to seek an alternative source?
Overtrust Question
Q3How often do you use AI outputs directly without checking them, assuming the system is accurate?
Overtrust Question
Q4How frequently do you present AI-generated work as your own without disclosing it came from AI?
Results — SELF Perspective
Trust Calibration Assessment
Relational Trust
Institutional Trust
Self-Trust
Task Trust
Overall Trust Calibration
Avg Undertrust
—
Avg Overtrust
—
The ideal trust score is calibrated — both Undertrust and Overtrust are low (≤2.5). If Undertrust is high: resistance, avoidance, shadow systems. If Overtrust is high: blind compliance, automation bias, inadequate verification. If both are high: confusion and organizational fragmentation around AI.
WHAT NEXT?