Analysis · Technology
ChatGPT Felt Less Creepy, Yet Students Reported Using It Less
By AWEI · AI-compiled · Published · Analysis prepared · 1 source · doi.org
Feeling more comfortable with AI does not necessarily mean using it more. In a 2024 study by Athanasios Polyportis at Delft University of Technology, students reported less unease about ChatGPT after eight months, alongside lower educational usage scores. Trust and perceived control over using the tool also declined.
The study followed 222 students enrolled in Dutch higher education who completed two surveys eight months apart. All had previously used ChatGPT for educational activities. Identical questionnaires measured reported use, trust, unsettling feelings, and whether students felt they had the knowledge, resources and control needed to use it.
The average usage score fell from 3.67 to 3.05 on a seven-point agreement scale, a statistically significant decline. These numbers describe agreement with statements about using ChatGPT for learning and assignments. They are not counts of sessions, hours spent or completed tasks—a crucial distinction when interpreting adoption evidence.
Polyportis then calculated each participant’s changes and used regression to examine how they related. Changes in trust and perceived control were positively associated with changes in reported use; changes in unease were negatively associated. Including age and gender, the model accounted for about a quarter of the variation in usage changes. Falling unease across the group therefore coexisted with falling overall use.
These associations do not establish what caused the changes. The paper suggests unreliable answers or institutional restrictions as possible explanations, but did not test those mechanisms. Self-reports, only two observations and the loss of participants between surveys limit interpretation. The findings concern previously experienced student users in the Netherlands, limiting transfer to other settings.
For teams evaluating AI adoption, the useful lesson is to revisit the same users and examine trust, discomfort and practical constraints separately. This study supports asking how engagement changes over time. It does not establish whether more use improves learning, saves time or delivers organizational value.
Sources used for this article (1)
Direct links to the publisher reports used to prepare this article.
- Source 1
- A longitudinal study on artificial intelligence adoption: understanding the drivers of ChatGPT usage behavior change in higher education Athanasios Polyportis
