Generative AI isn’t just reshaping tech. It’s reshaping how we think about learning.

I recently read (and now sadly can’t find the original to link!) a piece arguing that GenAI will mainstream reinforcement learning (RL). The TL;DR from that piece:

GenAI is pushing RL from a niche technique into the mainstream because LLMs need constant, structured feedback to stay useful. The future of these systems will focus on rewarding sustained behavior change—not just clicks—requiring teams to master reward design, long-term outcomes, and multi-agent feedback loops.

That framing is exactly what we need in L&D.

For years, we’ve measured success by completions and quizzes. What if we designed learning systems that truly:
-Reinforce the right on-the-job habits
-Tune difficulty and practice in real time
-Measure impact weeks or months later, not just at the end of a course (I really want valuable Kirkpatrick L3s and L4s!!)

This “living feedback loop” is a goal many orgs aspire to, but budget woes, tech debt (read: the cost of patchwork systems that are hard to modernize), and compliance pressures often drag us back to static training. GenAI may make it harder to ignore that gap.

Mainstreaming RL in L&D could mean shaping behavior over time, not just delivering content once.

Where are you seeing this future take shape or stall out?