Building AI Literacy in Risk-Averse Environments
← Back to Building Analytics Excellence
A Framework for Change Management and Technology Adoption
How do you build organizational capability when the dominant emotion is fear? I developed and tested this framework at Michigan State University, where restrictive AI policies created an environment where employees were paralyzed by uncertainty. The principles apply to any organization struggling to help people adapt to rapid technological change.
The Challenge
In 2025, AI tools were transforming knowledge work; yet, many organizations struggled to help their workforces adapt. At Michigan State University, restrictive policies created an environment where employees were hesitant to use AI, even for appropriate applications openly. Staff who could benefit from productivity gains were paralyzed by uncertainty about what was permitted. This presented a complex change management challenge: How do you build capability when the dominant emotion is fear?
The Framework
I developed a dual-track approach designed to reduce fear, build confidence, and create sustainable peer learning:
AI Office Hours: Low-Stakes Learning
Twice-weekly drop-in sessions prioritizing psychological safety; no registration, no judgment, just questions and real-time problem-solving.
AI Tea: Peer Knowledge Sharing
Monthly showcases where colleagues demonstrated their AI applications. Seeing trusted peers successfully use AI reduced perceived risk and provided contextual learning in familiar work environments.
Curated Resources
A comprehensive resource hub with scaffolded entry points, multiple learning modalities, and ongoing curation to address information overwhelm.
The Reality: When Good Design Meets Organizational Fear
After ten weeks, adoption remained modest. But this outcome itself provided valuable insights. Participation patterns revealed that institutional fear was functioning exactly as change management theory predicts: early adopters engaged immediately, while the larger majority remained hesitant despite available resources. Private conversations revealed significant ‘shadow usage,’ where staff used AI but were unwilling to discuss it openly, indicating that the fear of policy violations outweighed their curiosity about capability building.
Key Insights for Change Leaders
Infrastructure ≠ Adoption
Creating an excellent learning infrastructure is necessary, but it is also insufficient. When organizational culture emphasizes risk avoidance, even the most accessible resources cannot overcome psychological barriers.
Shadow Usage as Leading Indicator
Gaps between private usage and public acknowledgment indicate organizational trust deficits. Change leaders should treat shadow usage not as noncompliance but as evidence that current structures don’t serve staff needs, and as a pool of potential advocates once psychological safety improves.
Transferable Applications
This framework, developed within the constraints of higher education, translates to any organization navigating rapid technological change. Athletic departments adopting analytics face parallel challenges: varying technical capabilities, resistance from traditional mindsets, and siloed departments. Universities consistently struggle to strike a balance between innovation and risk management. The core principles – creating psychological safety, building peer networks, and measuring what matters – apply whether the technology is AI, advanced analytics, or emerging digital tools.
The Takeaway
Successful technology adoption requires equal attention to technical infrastructure and human factors. Organizations that thrive in rapidly changing landscapes aren’t those with the most sophisticated tools; instead, they are those that adapt quickly to new challenges. They’re those that help their people overcome fear, build competence incrementally, and create cultures where experimentation is celebrated rather than punished.
The lackluster adoption numbers weren’t a failure; they were data. They revealed precisely where organizational culture was constraining progress and what interventions would be needed. That diagnostic insight, and the willingness to operate in ambiguous environments, may be the most valuable outcome of all.