Abstract

In The Ends of Tests, Cope, Kalantzis, and Saini propose a transformative vision for education in the era of Generative AI. Moving beyond the limitations of traditional assessments—especially multiple-choice and time-limited essays—they advocate for AI-integrated, formative learning environments that prioritize deep understanding over rote recall. Central to their argument is the concept of cybersocial learning, where educators curate AI systems using rubric agents, knowledge bases, and contextual analytics to scaffold learner thinking in real time. This reconfigures the teacher’s role: not diminished by AI, but amplified through new pedagogical tools. The authors call for education systems to abandon superficial summative assessments in favor of dynamic, dialogic, and multimodal evaluations embedded in everyday learning. Importantly, this model aims to redress structural inequalities by personalizing feedback within each learner’s “zone of proximal knowledge.” Rather than automating outdated systems, the paper imagines AI as a medium for epistemic justice, pedagogical renewal, and educational equity at scale.

From the Paper:

The dangers of Generative AI are myriad. By now they have been thoroughly documented and widely acknowledged. Its extensive hazards and fast moving evolution require caution (Cope and Kalantzis 2023). However, this short paper does not focus on the well-documented negatives. Rather, it addresses the positive potentials and new opportunities for education. These, we will argue, can only be adequately realized in carefully curated AI education spaces.

The particular focus in this paper will be on the potential AI offers to transform assessment, and upstream from that, pedagogy. We make our case in broad brushstrokes, but also with brief reference at the end of the paper to a research and development initiative of ours applying Generative AI to education, CyberScholar. This exemplifies an alternative to legacy assessment systems while mitigating the worst of AI’s deficiencies, providing evidence of what is practically possible now and signposting possible futures.

The paper is organized around what we consider to be the three great potentials for Generative AI in education.

1. Careful combination of Contextual AI and Generative AI could bring about a quantum leap in pedagogy and assessment, moving learners from superficial content learning to deep learning.
2. AI can magnify the effectiveness of teachers, raise the level of their professionalism, and enhance the quality of their professional lives—but this will be quite a different job for which they will need to become AI-proficient.
3. AI may open a path to epistemic justice for all, in a world where, until now, education has been complicit in the systemic reproduction of wider social inequalities.

* Preprint, forthcoming as: Cope, Bill, Mary Kalantzis and Akash Kumar Saini, "The Ends of Tests: Possibilities for Transformative Assessment and Learning with Generative AI,” in AI and the Futures of Education: Disruptions, Dilemmas and Directions, Paris: UNESCO, 2025 

* Video here: https://mediaspace.illinois.edu/media/t/1_6j0sx6jr

* Full text here: 

Ends of Tests