In his free time, Alyn Kinney likes to watch old instructional films. The videos range from a 1940s video on “the anatomy of a bullet” to a 1980s “cringy rap video about how to fill up a soda cup.”
Kinney, who is a Sr. Manager of L&D at T-Mobile and the author of the “Nerd Out” L&D Substack, developed this hobby because he’s fascinated by the journey of learning. “I’m fascinated by where L&D came from and where we’re headed. These videos are a way to mine L&D’s past for the things that actually worked better than our current trends.”
This perspective is especially valuable right now, with AI dominating the conversation in L&D as a tactic to make training shorter, cheaper, and more frictionless. Kinney, who sits at the forefront of AI adoption in L&D, sees a more complex reality than just “simplifying at all costs.” As AI makes content creation easier, Kinney sees a clear advantage in moving back toward deeper, higher-fidelity learning: more in-person experiences, more coaching, and in some cases more ambitious internal learning ecosystems.
Across more than 40 interviews with L&D leaders, we’ve seen enormous disparity in AI adoption. Some organizations barely allow AI tools in the door. Others run advanced experiments, hold AI use case competitions, and are piloting AI-enabled learning programs.
Kinney’s perspective stands out across all 40 interviews because he is both pro-technology and deeply skeptical of bad thinking. He is not anti-AI. He is anti-buzzword, anti-slop, and suspicious of any trend that makes learning look “easy.”
Some of the Best AI Adoption Is ‘Invisible’
Kinney is wary of blanket enthusiasm for AI. “You can’t just use AI at every single learning point,” he said. In his own work, he uses AI for research, planning, and setup, but he is far more cautious about using it in employee-facing learning. “A lot of my most effective AI use cases are invisible and behind the scenes,” he said. “For example, I find that it’s tough to use avatar videos in learning. As soon as people feel that uncanny valley of the generated image, they get turned off. Then, they're thinking about AI and not the topic at hand.”
For Kinney, AI most often speeds up or deepens his work from “behind the scenes,” and this leaves him more time to make the human-facing experience cleaner and more intentional.
Cognitive Offloading Creates Cognitive Debt
Kinney pointed out that cognitive offloading is not new—or specific to AI for that matter. It’s also not always bad. In many roles, people shouldn’t memorize everything. A call center agent does not need to memorize every product feature if that information changes constantly and is stored in a convenient way. In those cases, mental offloading is efficient.
The issue is that AI makes offloading so easy and so smooth that workers can mistake project completion for understanding. When you outsource the mental work that leads to judgment, fluency, and recall, you go into what Kinney calls “cognitive debt.” “You could be in a board meeting, landing a huge deal with a team of lawyers. If you don't know your stuff, if you've been using AI to kind of outsource your thinking, you aren't going to be able to call on that knowledge in your moment of need,” he shared as an example.
That is where L&D has to get more intentional. Kinney argued that if something truly matters, training cannot remove all the friction. This could mean designing more retrieval practice, more conversation, more reflection, or more learner effort into the program. It could also mean teaching people to use AI differently. Instead of asking AI to do the thinking, ask it to sharpen your own thinking. “Use AI as more of a sparring ground,” Kinney said. “Ask it to tutor you.”
From Your LMS to Their LLM
“People’s natural instincts are moving away from the LMS and toward the LLM,” Kinney said. His point being that people are no longer waiting for a course catalog or a formal module. They’re asking the tool right in front of them.
This creates a huge moment of opportunity in L&D. What Josh Bersin has long described as learning “on-demand” and “in the flow of work” may finally become more real in the AI era.
For Kinney, that means less emphasis on generic, off-the-shelf content and more emphasis on contextual interventions. Guidance can appear closer to the task itself. In compliance-heavy environments, for example, guidance can surface closer to the risk and at the point of decision.
This shift could also change the areas that L&D spends its time. If AI increasingly handles the first layer of information-seeking, then learning teams have more room to focus on coaching, context, and personalization.
AI Is Raising The Bar For Quality
Kinney has a brutally useful test for L&D teams: “Would my learners pay money to take this training?”
“You have to respect learners’ time,” he said. And in the age of generative AI, respecting learners’ time demands something more valuable than it used to. If an LLM can already generate a passable overview of a topic in seconds, then training needs to offer much more than that LLM. The training has to be clearer, more contextual, more credible, and more useful than the cheaper alternative.
One Thing About L&D’s Future Is Certain: L&D Must Solve Problems, Not Just Produce Content
Kinney believes that above all else L&D needs to stop behaving like a content factory and start acting more like a problem-solving discipline. He likes to borrow from product management and ask, “What’s the problem that I’m trying to solve here?”
The L&D leaders who thrive will be the ones who use AI in the moments where speed matters but add friction when mastery matters; and those who stay focused on the right problems, at the right moment, in the right way.
Like those old instructional films Kinney watches on YouTube, many of today’s learning fashions will someday look dated. The lasting advantage will belong to the teams who understood the true needs behind their company’s learning.
