21/09/2026
AI can make performance look stronger — without making understanding deeper.
That’s the risk behind cognitive bypass: when AI does the reasoning a learner was supposed to practise.
A polished answer can look like competence. But if the learner can’t evaluate the output, explain the reasoning behind it, or perform without AI support, the capability gap is still there.
As Dr Lydia Elliott explores in The Learning Guild, the challenge for L&D isn’t simply whether employees use AI. It’s whether learning experiences still require people to think, judge and understand.
The goal shouldn’t be to remove every difficulty from learning. Some productive friction matters.
Design learning that asks people to:
🧠 reason before prompting
⚖️ evaluate AI-generated recommendations
🔍 question assumptions and errors
📈 demonstrate capability beyond the final output
AI should support learning — not quietly replace the thinking that creates expertise.
How is your organisation making sure AI improves capability rather than creating the illusion of it?