Weakness in Focus

The End of Assessment as We Know It: GenAI, Inequality and the Future of Knowing by Mike Perkins and Jasper Roe
UNESCO Report, AI and the Future of Education: Disruptions, Dilemmas and Directions (2025)

The ability of AI systems to create both complex and human-sounding products (across all mediums) has made many of our fundamental assumptions regarding originality and authorship unworkable. For years, assessments have been used to surface an individual's level of understanding through independent tasks, particularly those which could be taken home. AI can now assist with, if not accomplish, these types of assessments with little to no detection. Perkins and Roe suggest that this situation is not a new crisis; instead, it is a moment that brings longstanding weaknesses into focus. The structure of assessment systems (which have always allowed for gamed student responses) is being tested at scale. Additionally, a significant divide is growing: digitally well-equipped schools are now designing assessments based upon collaboration, ethical decision-making, and human-based abilities, while digitally disadvantaged schools are likely to continue using high stakes, traditional exams. For school leaders, the challenges surrounding the impact of AI-generated content on the development of assessments will be both technological and strategic. Decisions about assessment design will need to consider issues related to equity, accessibility, and institutional values. Since AI-generated content can produce answers to questions posed in assessments, school leaders will need to rethink how they define and measure learning. This work includes creating transparent processes for evaluating evidence of student thinking, while assuring that the new forms of assessing student knowledge do not further widen the existing inequalities among schools.