Structure AI-assisted grading to shift TA effort from routine low-judgment decisions toward conceptually rich responses¶

Strategy · All strategies
Evidence · no claims cited

Description¶

The article recommends that AI-assisted grading be structured so that teaching assistants spend their time on the responses that are most conceptually rich or require careful interpretation, rather than on routine, low-judgment decisions. This preserves the pedagogical benefit that evaluating ambiguous or challenging cases deepens TAs' content knowledge and pedagogical judgment, while reducing overall grading load and freeing TA time for other instructional and scholarly responsibilities.

Design Implications¶

Context¶

Requirements¶

• An AI pre-grading pass with confidence filtering must identify which responses need careful human interpretation

Constraints¶

• The article notes encoding all human-discretion contingencies (partial credit for serious attempts, lenient error-propagation handling) in a rubric for an exam taken by several hundred students would be impractical

Target Learners¶

• teaching assistants grading large-enrollment university science courses

Target Learning Goals¶

• deepening TA content knowledge and pedagogical judgment through grading of ambiguous cases

Related Strategies¶

• Allocate human and AI effort by dimension type: automate explicit-criteria dimensions, retain human review for pedagogical judgment
• Use context-rich multifaceted problems with TA scaffolding and group exams to create epistemic doubt in large introductory courses

Examples¶

-

Key Sources¶

• Jan Cvengros and Gerd Kortemeyer. (2026). Assisting the grading of a handwritten general chemistry exam with artificial intelligence. Discover Artificial Intelligence. https://doi.org/10.1007/s44163-026-01606-4