Grading methods and grade inflation in the AI era
The project involves systematically analysing grade data at department and school level – assessment formats and institutional factors, including AI use in learning and assessment – to generate evidence-based insights to guide educational development and policy decisions.
Project context
The current situation is characterised by the ever-increasing use of AI, together with factors such as grade inflation and inconsistent grading practices. This situation affects everyone at KTH. Students are directly affected by their grades and need clear rules regarding the use of AI, as well as transparency in grading. Teachers are responsible for grading and need to adapt their teaching and supervision of AI use to the assessment methods applied. School management sets guidelines and is responsible for ensuring that grading methods are fair and consistent, but the basis for these decisions may currently be insufficient.
A quantitative analysis is therefore needed to examine how different factors collectively influence grades and whether and to what extent grade inflation exists. Examples of such factors include assessment methods (such as the effects of courses based solely on projects or assignments), differences in grading practices, pedagogical traditions, structural changes (such as course mergers and larger student groups), and the increasing use of AI, etc.
Purpose (outcome)
The purpose of the project is to generate evidence-based insights that can:
- Guide pedagogical development and policy decisions
- Promote consistent and transparent assessment standards across the university.
This will be done through a systematic analysis of grade data at the department and school levels, assessment formats, and institutional factors.
The project complements ongoing work on pedagogical development, assessment quality, and the responsible use of AI. It also contributes to a more systematic understanding of grading outcomes at the school level.
Project results (output)
The project will deliver the following concrete results:
- Grade analysis report: A quantitative summary of five years of course grades* showing trends and correlations (e.g. grade inflation and links to grading methods).
- Key factors: An inventory of pedagogical, structural, and AI-related influencing factors.
- Case studies: 5 to 10 detailed course analyses of selected courses with different grading methods, as well as benchmarking against the pre-AI era.
- Recommendations: Evidence-based guidance for fair and consistent grading in the AI era.
- Scalable methodology: A framework for school-level analysis that can be applied across KTH.
* This does not include all courses at KTH but is limited to selected courses at Bachelor’s and Master’s level in order to create a representative selection.
Time plan
Start date: 2026-08-01
End date: 2027-07-31
Project documentation
If you have a KTH ID, you can read the documents when logged in:
- Poster: The project’s first poster (in Swedish) will be available at the ”Storträffen” meetup autumn 2026
- Project directive (Swedish): Projektdirektiv_FrU26_2614-EECS_BetygsInflation.pdf