AI Grading: Higher Ed's Existential Threat?

Educational institutions are facing a critical juncture regarding the use of generative AI in marking student assignments. Senior lecturer Armin Alimardani highlights concerns that AI-driven grading could pose an existential threat to higher education by creating a 'slop cycle' of AI-generated submissions and AI-based assessments. This risks devaluing degrees and diminishing student motivation if feedback quality declines.

Several Australian universities, including Western Sydney University, the University of Newcastle, Deakin University, RMIT, and the University of Adelaide, have introduced qualified allowances for AI in assessment. However, these often emphasize that academic staff retain ultimate responsibility for grading and feedback, with students sometimes having opt-out rights. Critics worry that over-reliance on algorithms could replace human academic judgment and lead to 'verification drift', where human oversight becomes less vigilant.

The potential benefits of AI in marking, such as assisting overworked academics, are being weighed against significant risks. Some institutions, like the University of NSW, University of Melbourne, and University of Sydney, are drawing a firm line against AI in grading altogether. Meanwhile, universities like Monash and Wollongong are exploring AI integration into teaching and assessment to support learning, though not yet for direct marking.