Moderate evidence
More
Australian students cite AI's help with study and efficiency, and flag inaccurate outputs
Study title: Comparing the positive and negative impacts of AI: A large-scale study of student perspectives
- Evidence type
- Survey
- Sample
- n = 2,802
- Sources
- 1 sourcePeer-reviewed
- Published
- 2 min read
- Truvace record
- TRV-2026-2999cite or check it
Study title: Comparing the positive and negative impacts of AI: A large-scale study of student perspectives
Researchers analysed open-ended comments from 2,802 participants, drawn from a survey of 8,021 students at four Australian universities, to see how students weigh the positive and negative effects of generative AI on their studies.
Read the study abstract
Higher education students are incorporating generative artificial intelligence into many aspects of their studies, yet the ways they judge its positive and negative impacts remain poorly understood. Using qualitative content analysis supported by code-frequency analysis, this paper examines open-ended comments from 2,802 participants drawn from a survey of 8,021 students across four Australian universities. Reported impacts clustered around four themes: study practices, quality, efficiency, and institutional pra…

57Gain side
How the two sides scoreClose call
58Problem side
The gain
In an analysis of 2,802 open-ended comments from a survey of 8,021 students at four Australian universities, positive AI impacts were coded more often than negative ones, mainly around students' own study practices and efficiency.
The problem
In the same-university survey of 8,021 students, analysis of 2,802 comments found negative responses centred on inaccurate AI outputs, and the few comments on institutional teaching and assessment practices were mostly negative.
A fixed formula sets both scores from the evidence and its sources. Reader votes don’t move them.
What the scores mean
G is the gain side, out of 100. It goes up when the gain is specific, when it measurably helped people and when the sources behind it are strong.
P is the problem side, out of 100. It goes up when the problem is specific, when it measurably hurt people and when the sources behind it are strong.
In brief
Comments clustered around four themes: study practices, quality, efficiency and institutional practices. Positive impacts were coded more often overall, especially where students described AI supporting their own study practices and efficiency.
Students did not simply embrace or reject AI. The authors say they made situated judgements shaped by concerns about accuracy and trust and by how much agency and control they kept over its use. All findings are based on what students said, not on measured learning results.
Main points
- The analysis covered open-ended comments from 2,802 participants, taken from a survey of 8,021 students.
- Students were drawn from four Australian universities.
- Comments fell into four themes: study practices, quality, efficiency and institutional practices.
- Positive impacts were coded more often than negative ones, particularly for study practices and efficiency.
- Negative comments centred on the inaccuracy of AI outputs.
- Few students commented on institutional teaching and assessment practices, but those comments were mostly negative.
- Students said AI helps them begin, continue, refine and complete academic work.
The full story
The paper uses qualitative content analysis supported by code-frequency analysis of free-text survey comments. The 2,802 commenters come from a larger survey of 8,021 students at four Australian universities.
Four themes organised the reported impacts: study practices, quality, efficiency and institutional practices. Positive coding dominated overall, especially where students described using AI to support their own study practices and efficiency.
Accuracy was the main source of negative comments. Comments about institutional teaching and assessment were relatively few but predominantly negative. The authors suggest students value AI for helping them start, continue, refine and complete work, and that institutions should engage with the conditions under which students find AI useful or problematic.
What this doesn’t fix
The findings rest on students' own open-ended comments, and only 2,802 of 8,021 survey respondents wrote comments, so they reflect perceptions rather than measured learning outcomes.
Sources
1 on file- ASCILITE Publicationsdoi:10.65106/apubs.2026.3397
Talk it over
Positive comments were coded more often than negative ones, mostly about study practices and efficiency, while negative comments centred on inaccurate AI outputs. How would you weigh those against each other?
Australian students cite AI's help with study and efficiency, and flag inaccurate outputshttps://truvace.com/item/comparing-the-positive-and-negative-impacts-of-ai-a-large-scale-study-of-student
Positive comments were coded more often than negative ones, mostly about study practices and efficiency, while negative comments centred on inaccurate AI outputs. How would you weigh those against each other?
Answer in the threadOnly 2,802 of the 8,021 surveyed students at four Australian universities wrote comments. How might the students who did not write comments see AI differently?
Answer in the threadFew students commented on institutional teaching and assessment practices, and those comments were mostly negative. What finding would tell you whether that small group reflects a wider view?
Answer in the thread
If you have more time
Students said AI helps them begin, continue, refine and complete academic work. What would you ask students at a university you know to learn how much control they feel they keep?
Answer in the threadThe comments describe perceptions, not learning results. What measured outcome could be compared with what students say about accuracy and trust?
Answer in the thread
ace

What readers say