A university allows essays to be graded by an AI system.
The model was trained on thousands of previously top-scoring papers.
After implementation:
• Unconventional texts receive lower grades.
• Traditionally structured essays perform better.
The AI identifies structural similarity to past top papers as a quality indicator.
Experimental formats naturally deviate from this pattern.
Question:
Is the system rewarding quality — or conformity?
Solution follows tomorrow.