From creation to transformation: Evidence of Learning in the AI Era
Artificial intelligence is changing not only how learners complete academic work, but how they begin it. For generations, learning often started with cognitive friction: the uncertainty of the first sentence, the struggle to organize ideas, and the challenge of moving from ambiguity to momentum. Today, intelligent tools increasingly reduce that initiation burden by supporting brainstorming, outlining, information gathering, drafting, and revision from the earliest moments of the learning process.
This shift carries significant implications for how institutions interpret learner work, as learners no longer begin from the same blank-page conditions that have historically shaped academic work. Final essays, projects, and examinations offer diminishing visibility into what educators most seek to understand: how learners think, decide, adapt, and apply knowledge with purpose. Completed artifacts may still demonstrate learning, but they no longer serve as sufficient standalone indicators of human mastery.
Higher education must therefore shift from artifact-centered assessment to evidence-centered models that make learning visible over time. Building on the conceptual foundation established in Reframing Bloom’s for the Age of AI (Clark, 2025), transformation represents the most credible indicator of advanced learning in AI-mediated environments. At this level, learning is evidenced through the learner’s ability to exercise judgment, make contextual decisions, engage ethical reasoning, and adapt their thinking throughout the learning process.
A practical response to this shift is found in the Transform Evidence Arc (TEA) which has been informed by ongoing conversations with institutional leaders, instructors, and academic teams. The TEA is a longitudinal framework capturing observable evidence of human capability across six interconnected stages: Intent, Exploration, Judgment, Application, Reflection, and Iteration. Rather than relying on a single submission as proof of competence, the Arc enables institutions and instructors to evaluate the movement of learning over time, revealing the distinctly human decisions and adaptations that intelligent tools alone cannot provide.
To preserve assessment credibility, however, developmental evidence alone is not enough. Embedded throughout the arc are Human Validation Checkpoints, intentional moments in which learners must explain, defend, articulate, or transfer their understanding in ways that confirm authentic command of the work. Together, the Transform Evidence Arc and Human Validation Checkpoints provide a more credible and sustainable model for evaluating learner readiness in educational environments where AI is now part of the process rather than outside of it.