It's been about three years since automated content generation became good enough that L&D teams started seriously asking what it should and shouldn't do. The honest answer in 2026 is: a lot more than the cautious early adopters predicted, and a lot less than the marketing copy claims.
What automated authoring is genuinely good at
- Structuring content from a brief into chapters, knowledge checks, and a final assessment.
- Drafting plausible distractors for multiple-choice questions.
- Producing competent synthetic narration for non-broadcast contexts.
- Translation and localisation drafts for multi-language modules.
What automated authoring is bad at
- Knowing your organisational context.
- High-stakes correctness in regulated industries.
- Replacing instructional design judgement.