When Your Team Stops Experimenting, Check What You Know About AI
Your team’s willingness to try new tools depends heavily on team dynamics. When a leader visibly struggles with AI, their people quietly shelve their own experiments. Psychological safety around technology doesn’t appear by accident, it follows the leader’s competence and confidence.
Forero et al., in ‘The Role Of Leader’s AI Literacy When Appropriating AI’, published in the Journal of the Association for Information Systems in 2026, studied exactly this chain of cause and effect.
How Leader AI Literacy Shapes Team Dynamics in Practice
The researchers wanted to understand how a leader’s grasp of AI shapes the way teams actually adopt AI tools at work. They conducted a multiple-case study across two organisations: a Colombian financial cooperative and a German insurance company. The method combined interviews, observations, and document analysis across both sites.
The findings are pointed. Leaders with strong AI literacy actively modelled experimentation. Their teams followed suit, treating errors as data rather than failures. Where leaders lacked AI literacy, teams defaulted to surface-level use, clicking through features without adapting workflows. One telling pattern emerged across both sites. Teams mirrored their leader’s comfort level almost exactly. Leaders who asked questions about AI outputs in meetings gave their people permission to do the same. Leaders who stayed silent on AI limitations signalled that silence was the safer choice.
Turning Research Into Coaching Conversations
This finding has direct consequences for how you develop senior leaders. Start by auditing what your leaders actually know about the AI tools their teams use. Not the marketing pitch, the mechanics. Can they explain why a model might produce a confident but wrong answer? Can they describe one failure mode their team should watch for?
Amazon’s approach in its AWS divisions offers a useful model. Team leaders are expected to demo tools before rollout, exposing their own learning curve publicly. That visible imperfection builds permission for the team to experiment without fear of looking incompetent. Contrast that with the pattern seen at many European banks in 2024 and 2025. Leaders delegated all AI adoption to IT teams. Front-line staff read that as a signal: this technology is not for us to question. Adoption stalled, and so did the benefits.
Coach your leaders to narrate their own AI learning out loud. A simple habit works well. Before a team meeting, a leader spends two minutes sharing one thing they got wrong with an AI tool that week. That single behaviour reshapes team dynamics more than any formal training programme.
What This Research Cannot Tell You
The study draws on two organisations in two countries. Both are financial services firms, which limits how far you can generalise the findings to manufacturing, healthcare, or professional services. The case study method captures depth, not scale.
I keep returning to the image of a leader staying quiet during an AI demo because they don’t want to expose a gap in their knowledge. It’s a completely understandable impulse. But that silence travels. It tells every person in the room that not knowing is a risk worth hiding. The research from Forero and colleagues makes the cost of that silence concrete.
If you lead a team that’s underusing AI right now, I’d ask you this: when did you last admit, out loud and in front of your people, that you didn’t understand something the tool produced?
Image: Photo by Campaign Creators on Unsplash
Explore more insights at Ariston Institute.
