Is AI is Making Your Experts “Smarter” AND Destroying Pathways for Development?

During the December LILA gathering, Matt Bean shared his research on Skill Code.

The paradox few are talking about is that while AI delivers 10-15% productivity gains for 80% of your workforce, it’s simultaneously dismantling the expert-novice relationships that have built human capability for centuries. This means that new employees must no longer have development opportunities creating a situation where they are in effect entering the organization at the mid-career level with little or no experience or pathways for development..

Matt Bean found that in robotic surgery, resident engagement dropped from 4-4.5 hours of active learning per procedure to just 10-15 minutes. When automation makes experts more self-sufficient, they naturally capture the most valuable, complex work leaving novices sidelined without the hands-on practice that builds genuine skill.

The result? “A multitrillion-dollar skills gap. And we have a 3-5 year window to fix it.” Matt Beane

The Shadow Learner Signal

Matt stated that 1 in 8-10 learners is now a “shadow learner,” someone bending or breaking rules to access the developmental experiences your formal systems no longer provide. They’re the canary in the coal mine.

As one leader put it: “Many skills should die. Many must live. How can we know the difference?

The question isn’t whether to deploy AI. It’s whether you’re designing it to amplify or replace the three fundamentals of skill development: Challenge (stretching at the right edge), Complexity (wrestling with real, messy problems), and Connection (learning together, not in isolation).

What Leaders Must Do Now:

✓ Audit every AI deployment: For each task you automate, identify which novice lost a learning opportunity—then redesign to preserve it.

✓ Set a dual mandate: Every technology investment must demonstrate both productivity gains and measurable skill development. Reject proposals that optimize only one.

✓ Hunt for the teams where juniors are thriving with AI. Document what makes it work. Replicate it everywhere.

✓ Interview your “shadow learners.” Ask what they’re fighting to preserve—then build those conditions into formal workflows so learning doesn’t require heroics.

The challenge before you: Can you maximize productivity and skill development—or will you optimize for one at the cost of the other?

What’s one AI deployment in your organization you need to audit right now? Share in the comments—let’s figure this out together.