Every week, another vendor promises that AI will "automatically generate" your training content. Click a button, get a course. It sounds appealing - until you ask what the learning objectives are, whether the content maps to a performance gap, or how you are measuring whether anyone actually learned anything.
Here is the thing: AI is genuinely useful in instructional design. I use it daily. But it is a production accelerant, not a replacement for the thinking that makes learning stick.
The discipline of instructional design exists because humans do not learn the way we often assume. Cognitive load matters. Prior knowledge matters. The sequence of information matters. Feedback loops matter. None of these considerations disappear just because you can generate 20 slides in 30 seconds.
What changes is the stakes. When content generation gets cheaper and faster, the differentiator becomes the quality of the design decisions behind it. Organizations that invest in genuine instructional design expertise will produce AI-assisted content that works. Everyone else will produce content that looks finished but does not move the needle.
That is the opportunity - and the risk - right now.