Education

Universities cannot grade their way out of AI

A higher-education analysis said universities need to rethink assessment rather than rely solely on AI detection or AI-resistant assignments.

A report on higher education argued that universities cannot address generative AI simply by trying to design assessments that AI cannot complete. The discussion followed controversy in the UK over a proposed change that was paused until September 2027 for redrafting with the university IT body UCISA.

The issue reflects the difficulty of maintaining traditional assessment models as generative and agentic AI systems improve. Detection-based approaches may become less reliable as students gain access to more capable tools.

The report points toward broader changes in assessment, including greater attention to process, oral explanation, reflection, and evidence of independent thinking. Such approaches assess how students reached an answer rather than relying only on the final submitted product.

The development matters because assessment is a core operational challenge for universities adopting AI. Institutions must balance legitimate AI-assisted learning with academic integrity, while ensuring that qualifications continue to demonstrate students’ own knowledge and capabilities.

Original source: AI Weekly

Related RAQI solutions

In-depth analysis on RAQI Insight

Work with RAQI