AI assistants are very good at removing the struggle from learning. That is the problem.
Cognitive scientists describe desirable difficulties: retrieval, spacing and effortful generation strengthen memory, while fluent reading and copying feel productive but leave little behind. An AI that produces the summary, the answer or the first draft removes exactly the friction that builds durable skill, especially for novices who have no schema against which to check the output.
The practical distinction is between AI as an answer engine and AI as a sparring partner. One compresses time; the other compresses thinking.
- Trainer: let learners attempt before the AI reveals anything.
- Instructional designer: put the attempt inside the task, not after it.
- L&D manager: measure retention a month later, not satisfaction scores.
- HR leader: ask which capability is quietly being deskilled.
The design question is not whether AI sits in the workflow, but where in the sequence it sits: before the attempt, or after it.
Comments 0
Sign in to like, save or comment.
No comments yet. Start the discussion.