News
AI in learning: choose the task before the tool
A new UNESCO study gives training centres a reason to ask what a digital assistant actually helps people learn.
On 9 September 2026, UNESCO IESALC presented in Paris a study of 200 higher education institutions in 19 Latin American and Caribbean countries. Among participants, 87% reported using AI and 26% having a formal strategy. These figures describe the participating institutions, not every university worldwide. Sources: launch announcement, study overview.
Something a training centre can try
What follows is an editorial proposal, not a finding or recommendation from the study’s authors. Start with one small task and decide beforehand how to check understanding without the assistant.
Consider this hypothetical example: on an entrepreneurship course, a learner asks AI to suggest questions about an order description. The learner then independently lists what to clarify with the customer: quantity, deadline and acceptance criteria. The teacher checks that list against the original description. A polished answer does not by itself show that someone has learned to handle an order.
Agree on four points before the session:
- Which actions the learner must complete independently.
- Who checks facts and corrects the assistant’s errors.
- Which information must stay out: fictional details without customer contacts are enough for the exercise.
- How to complete the same task without a paid subscription or a personal device.
Make the benefit observable
After the trial session, offer a similar task without AI. Check whether the learner can explain the solution in their own words and spot missing conditions. This small exercise cannot establish the technology’s effectiveness in general, but can help decide whether to continue this particular format. Discuss purchasing access after checking the learning benefit and the full cost for the group.