Continuing professional development: 83 per cent of candidates read AI-generated summaries, and 86 per cent are wary of them

A UPCEA survey of 1,057 adults shows that the mention of a school in an AI response now influences a list of preferences that has been reduced to just three schools.

UPCEA and Search Influence surveyed 1,057 adults aged 18 to 60 who are interested in continuing education or online training from 5 to 14 May, and published their results on 16 September. The search engine remains the starting point for 80% of them, but 54% also use AI tools in their decision, and 83% read the automatically generated overviews when they are displayed.

Two curves intersect. Trust in the institutions cited as sources in these responses is increasing — 64%, compared to 56% in 2025 — while mistrust of the tool itself is growing even faster: 86% of respondents express at least one reservation about the reliability of AI, compared to 68% a year earlier. Candidates are therefore using a system they do not trust, and still retain what it tells them.

The effect is mainly measured downstream. The majority of respondents now examine only a maximum of three institutions before applying. On such a short list, whether or not to be mentioned in a generated response becomes a discriminating factor, which the authors summarize by observing that a citation may be enough to make an institution part of the final selection.