Same Performance, Different Process: Epistemic Ownership in AI-Mediated Education

arXiv:2610.02731v1 Announce Type: new Abstract: Generative AI weakens the link between assessed performance and the cognitive processes that educational work is expected to reflect. This paper examines observable traces of those processes in a sample of 150 student-AI conversations. The analysis uses the concept of epistemic ownership as an interpretive framework, focusing on three dimensions that can become visible in interaction: Direction, Integration, and Evaluation. The results show that similar assessed performance can coexist with substantially different observable patterns of cognitive participation. Assignment scores therefore provide information about assessed outcomes but do not recover how cognitive work was distributed between student and AI during the interaction. The study provides an empirical basis for asking how assessment should incorporate evidence about cognitive process, rather than relying exclusively on the quality of the resulting product.

Study questions current assessment methods

This is a measured empirical finding about assessment validity gaps in AI-mediated learning, not a dramatic claim. It implies a real but moderate concern: current grading may not capture whether learning occurred, which has negative implications if uncorrected but is also actionable.

The study empirically demonstrates that AI can obscure student learning processes, even when performance metrics remain constant. This weakens the reliability of traditional assessments regarding actual student comprehension and effort. The implication is a potential degradation of educational integrity and the meaningfulness of qualifications.

The paper highlights a significant shift in educational assessment due to AI, suggesting a need for new methods. This could lead to more comprehensive evaluations of student learning.

SOURCE     arXiv cs.CY (Computers and Society) (tier 1)
   ↓
DOCUMENT   48a383b1-e331-4cde-8b9e-04b2f5271e31
           https://arxiv.org/abs/2610.02731
   ↓
EVIDENCE   2 extracted excerpts
   ↓
MODEL RUN  4 runs, methodology 1.0.1
   ↓
SCORE      -41  (Adverse)
   ↓
CONFIDENCE 74%