Take Control! Don’t Let AI Define You!
If you didn’t have boundaries and limitations (budget, technology, policy, culture, skills, etc.), would you do your job as a learning specialist the same way you do today? That’s the big question for AI-augmented learning tasks. We need to review the fundamental questions of why and how we work; who we are; and how we create value. The first obvious improvement with AI is always efficiency: reducing hours spent on content creation (measured in human-equivalent hours). That’s a start. Or maybe, a dead end?
In this article, I challenge you to step back from the trenches of rapid content development with AI. Let’s think about your current workflow! From the training period, through design, development, implementation, and possibly, measurement and evaluation.
In this article…
How Did We Get Here?
At that time, we had different teams that developed and delivered learning solutions: Instructional Designers (sometimes Instructional Designers), developers, operations, etc. IDs and ISDs were responsible for early needs assessment, and working with SMEs to create a storyboard for approval. Once the storyboard was approved by everyone (final_final_gold_reallyLastVersion_useThis.doc), development began.
Developers take an idea from paper and implement it with Flash or other fancy tools. Changing anything after development was a pain and could be a huge delay. Another group, L&D functions, will upload and check this package in the LMS. If this was an instructor-led approach (in-person or virtual), we would have to create guides, workbooks, slides, etc.
This was a really slow and fragmented process. The business began to demand faster response times. This is where fast eLearning tools come in. What Lectora, Storyline, and Captivate (just to name a few) have done in the eLearning industry has been blurring the lines between designers and developers. Everyone could build things. And everyone could design things. (Whether those “things” were working or not, that’s a different question.)
Those who became experts in both design and development had the advantage of creating magical learning experiences because they could simulate, iterate, and improve themselves, very quickly. They can translate SME and stakeholder ideas into communication quickly.
So far, however, we haven’t really changed the basics of our performance. We have recently improved our efficiency. As a store that produces products efficiently, so the inventory looks good.
From Warehouse to General Contractor: What AI is Really Changing in L&D
Speaking of warehouses. Let’s examine this analogy from the last twenty years of corporate L&D: a well-organized computer hardware store. Large catalog of pre-produced content. Users browse, select, use. We measure completions, registrations, and sometimes even information. We had a long debate about how to deliver (instructor-led vs. self-paced vs. blended) to offer live cashiers and self-pay. An end-of-course satisfaction survey was received. It’s friendly, fast, and almost entirely close to the point of impact. And then the user came out with his tools and parts, and the organization had absolutely no idea what he was building. Even if they build anything.
Content utilization and a warehouse mentality can make a learning organization look busy, but inefficient. The warehouse always weighed the wrong item. Walking. Catalog size. Termination. These are inventory metrics, not construction metrics. They tell you what left the shelf, not whether anything important was built.
Rapid eLearning technology was a purposeful technology to solve a specific problem: production cost and speed. It didn’t touch on the deeper problem, which was that no one was accountable for what happened after they left.
AI Is Not Technology
This is where the current era is truly different, and why it represents a paradigm shift rather than another technological productivity improvement.
Don’t treat AI as a new technology introduced to the warehouse!
AI doesn’t just help you build courses faster. It makes you face a question that the warehouse model never answered: if we can surface the right information at the exact time someone needs it, simulate a real conversation before the real thing happens, or provide personalized training based on real performance data, do we need to build courses the same way? Do we need to stock shelves? How would you redefine your work? Output? Your result? Your price?
The lesson was always to adjust. Workaround for the impossibility of having an expert available for every job every time when needed. The AI asks it’s impossible. Which means that the performance of AI-augmented learning may no longer be required in the same way for all problems. And the L&D function has to think about what it was trying to achieve all along. Change from warehouse to general contract.
How does a general contractor do it differently? This is not someone who keeps tools in stock, but someone who comes to the job site, assesses what exactly needs to be built, pulls the right resources at the right time, and is always accountable for whether the building stands when it is done. That’s a very different business model: different stakeholder relationships, different success metrics, different skills needed, different conversations to have.
The contractor has a problem, however
Here’s the part of the AI story that L&D professionals need to understand clearly, because getting it wrong has real consequences. An AI-powered human contractor can build and destroy with equal confidence at scale! If you let the AI explain it to you, it will.
The AI contractor is incredibly talented. It works at a speed that no one can match, it doesn’t tire, and it can handle a lot of information at once. But it has two important weaknesses that are almost exactly the same.
- First: it’s wrong to be more confident than it should be.
AI systems generate sound answers with consistent approval, whether those answers are accurate or not. In L&D terms, this means that AI-generated content can contain factual errors, outdated information, or subtly incorrect structures, delivered with the same smooth confidence as completely accurate content. You own the result. If you generate an AI flop, it’s on you. Time. - Second: do what you ask, prepared to please you.
AI systems are trained in ways that make them responsive and adaptive. When you ask for a lesson, you get a lesson and not a challenge to your thinking that the lesson is the right solution. When you ask leading questions, you get convincing answers. If your brief isn’t clear, the AI fills in the gaps with whatever seems satisfactory rather than flagging that the brief isn’t good enough. Set the co-worker/inventor rules: you have to explain how the AI works with you.
Together, these two trends can reinvent (at the speed and scale of AI) the same problem that the warehouse has always had. A customer who doesn’t know what they really need, combined with a system designed to fill the order rather than ask, produces more confident, well-packaged, and wrong answers very quickly. Drywall for everyone!
What This Means for the L&D Professional in an AI-Augmented Learning Environment
The general contractor metaphor clarifies what the new role of L&D for AI-enhanced learning really requires. A good contractor doesn’t just do briefs for a client. They back off when the summary is wrong. He said, “You asked for this wall here, but structures that carry loads don’t work like that.” They bring expert judgment that the client does not have and does not pay them to suppress.
This is precisely the skill set that L&D professionals need to develop in relation to AI. It’s not just informing, although informing is really important, but knowing when to investigate the result, when to pass on the recommendation, and when the AI’s definitive answer is built on faulty assumptions embedded in the question.
In other words, you will need to explain who you are and how you work for AI. Don’t let AI do it for you. Yes, you can take the back seat and enjoy the ride, but that won’t get you where you want to be.
How to Drive Coworking?
The L&D professionals who will thrive in AI-enhanced learning environments are those who hone three skills in particular.
- First is the rigor of the diagnosis
The ability to accurately identify what operational problem actually exists before reaching for a solution, resisting both organizational pressure to produce content and the willingness of AI to produce it when needed. - The second is an important assessment
Treating AI-generated content like a skilled first draft from a young colleague, not a finished product from an expert, and reviewing it with the same scrutiny you can apply to anything with real results. - The third is accountability for the result
That is to own the question of whether the intervention actually changed behavior, not just whether the content was delivered and used.
The Arehouse L&D specialist was responsible for the shelf. The L&D general contractor will be responsible for the structure.
Opportunity, Clearly Stated
None of this diminishes what AI makes possible. For the first time, L&D has tools that can bridge the gap between what employees always say they care about: performance, behavioral change, organizational strength. All this in measure! But that gap closes only if the people using AI understand its limitations as clearly as their strengths. A strong contractor without someone to check the plans is no better than slow progress. It’s a quick way to build something wrong. The profession now is not knowing AI. It is a judgmental mastery that AI cannot restore. Take control! Explain who you are! Don’t let AI do it for you.
