Studies: generative AI cannot grade essays accurately and tends to give higher marks
Cardiff University says generative AI cannot accurately grade higher-education essays, and another study found that AI tends to give higher marks than humans do. According to the reports, teachers are less likely to challenge grades given by AI.
Who is affected and how
- teacher
- AI grades should not be accepted without checking, because they are inaccurate and can skew upward.
- university student
- AI marking may be more favorable, but it is not necessarily accurate feedback.
- education policymaker
- Before AI grading is introduced, rules for human review and calibration must be set.
- edtech company
- Automated grading products must demonstrate accuracy and freedom from bias.
Sources
- GenAI cannot accurately mark essays in Higher Education - News - Cardiff Universitycardiff.ac.uk · Aug 24, 2026
- GenAI cannot accurately grade student essays in higher education, study findsphys.org · Aug 24, 2026
- GenAI Fails to Accurately Grade Student Essays, Study Findsnewsstreetdaily.com · Aug 24, 2026
- AI tends to mark students’ essays higher than humans – studytimes higher education · Aug 26, 2026
- AI Tends to Mark Students’ Essays Higher Than Humans, Study Shows - Inside Higher Edinsidehighered.com · Aug 27, 2026
- AI Tends to Mark Students’ Essays Higher Than Humans, Study Showsinside higher ed · Aug 27, 2026
- AI-generated grades less likely to be challenged by teachersthe educator · Aug 28, 2026
- AI-generated grades less likely to be challenged by teachers | The Educator K/12theeducatoronline.com · Aug 28, 2026
Machine-generated summary based on the titles and excerpts of the source articles. The facts come from the linked article.