IA on the job: does an easy-to-read text seem more credible?
Generative AI can quickly give a business memo or product presentation a professional finish. The managerial question is specific: how can we prevent ease of reading from standing in for evidence during approval?
What the literature establishes
Rolf Reber and Norbert Schwarz (1999, Consciousness and Cognition) showed experimentally that perceptual ease of reading could influence judgments about the truth of statements. Their manipulation concerned visual readability, not the quality of arguments. This finding establishes that a presentation feature can enter an assessment that should, in principle, concern the facts.
Processing ease as a cue
Adam Alter and Daniel Oppenheimer (2009, Personality and Social Psychology Review) review research on processing fluency, the subjective experience of ease with which information is processed. It can serve as a cue in different judgments, but its interpretation depends on context and on what that ease is attributed to. Applied to AI-assisted writing, this mechanism suggests a risk worth examining: mistaking ease of reading for a sign of strong documentary support.
The premature conclusion
One might conclude that texts should be made less polished to improve decisions. These studies do not justify that recommendation: clear presentation remains useful for understanding and discussing a case. The relevant distinction is not between polished and awkward writing, but between ease of understanding and justification of claims. (our executive and employee training programmes)
What the evidence does not establish
These articles do not measure the approval of AI-generated documents in companies in Ticino. A manipulation of visual readability cannot, on its own, predict the effect of professional writing on a banking or commercial decision. They offer a hypothesis that warrants attention, not an estimate of local risk or proof that an expert reader will necessarily be influenced.
A practical check in Lugano
As part of « IA on the job », an exercise in Lugano could use a fictional memo in Italian, tailored to a bank, fashion business, trading company or Italian-speaking family SME, without confidential data. Randomly assign participants to two layouts of the same text, both readable but one easier to scan, then ask them to rate its credibility from 1 to 5 before providing supporting documents. Record the difference between the average ratings; in a small group, it provides a starting point for discussing approval criteria, without demonstrating a general effect of AI. To go further: explore the AI on the Job training in Lugano, or browse our executive and employee training programmes in Switzerland.
In pictures: AI on the Job in Lugano
- ai on the job
- Lugano
- research
- training lugano