Most AI conversations in L&D start with capability: can it write the scenario, build the quiz, summarise the policy? That question is easy, and usually answered yes. The harder question comes next: which part of this work is actually mine?
A useful frame is a delegation ladder.
- Assist: AI suggests options, you choose.
- Draft: AI produces a first version, you rebuild.
- Check: AI reviews your work against criteria you set.
- Decide: AI recommends, a person owns the consequence.
Climbing the ladder is not about trust in the tool. It is about reversibility. A rough draft is cheap to reject; a design decision that shapes what hundreds of people practise is not.
For trainers, the ladder protects facilitation judgement. For instructional designers, it keeps task analysis and assessment alignment human. For L&D managers and HR leaders, it sets accountability before adoption rather than after.
Ask one question before the next prompt: if this output is wrong, who notices, and when?