AI is everywhere in the learning and development conversation right now, and the claims range from genuinely useful to wildly overblown. The truth sits in the middle. AI is a powerful tool that speeds up real parts of the work, and it is not a replacement for good design. Here is an honest look at where it helps and where it falls short.
Where AI genuinely helps
Speeding up the first draft
AI is good at producing a rough first draft of content, an outline or a set of quiz questions from source material. It does not produce the finished article, but it can get a designer past the blank page faster, leaving more time for the thinking that matters.
Voiceover and narration
AI-generated voices have improved a lot and are now good enough for many courses, especially drafts and internal training. They are quick, cheap, and easy to update when content changes, which is a real saving compared with re-recording a human voice artist.
Translation and localisation
For organisations with a multilingual workforce, AI makes translating and adapting content faster and more affordable, though a human check is still wise for nuance and accuracy.
Images and simple media
AI image generation can produce custom visuals quickly, which helps keep a course on-brand without a large illustration budget.
Personalisation and support
AI can help tailor what a learner sees next based on how they are doing, and AI tutors or chat assistants can answer questions in the flow of work. These are promising, and worth watching as they mature.
Where it is mostly hype
"Generate a whole course in minutes"
This is the big one. Tools that promise a complete course from a prompt produce exactly what you would expect: generic, click-through content that covers a topic without changing behaviour. It is the digital equivalent of the forgettable training people already ignore. Speed is not the same as quality.
Replacing instructional design
AI can generate content, but it does not know your people, your goals or what behaviour you are trying to change. The hard part of good training is not producing words, it is deciding what to teach, how to teach it, and how to make it stick. That is design, and it still needs a human.
Accuracy you can trust without checking
AI can be confidently wrong. For compliance, safety and anything where being right matters, AI output has to be checked by someone who knows the subject. Treating it as a finished source is a real risk.
What still needs a human
The judgement calls are where people stay essential. Understanding the real problem behind a training request, designing for behaviour change, choosing the right level of interactivity, getting the tone right for your culture, and making sure the content is accurate and accessible. AI assists with all of these. It replaces none of them.
How to use AI well in L&D
Treat AI as a fast assistant, not a finished product. Use it to speed up drafts, narration, translation and visuals, then apply human design and review to make it genuinely good. Be careful with sensitive data, and always have a subject-matter expert check anything where accuracy matters. Used this way, AI lets a skilled team do more, faster, without dropping the quality that makes training work.
At GloBug we use AI where it earns its place and rely on human design for everything that matters. You can read about how we approach the craft on our approach page.
Thinking about how AI fits into your learning plans? A free Health Check is a good place to talk it through.