For decades, educational assessment has functioned as a post-mortem examination rather than a tool for growth. Traditional tests deliver a static snapshot of a student’s abilities—too late to help them improve. Instead of guiding learning, these tests often simply rank students, failing to diagnose how they learn and mature. However, the rise of artificial intelligence (AI) may be changing this dynamic, bringing us closer to a vision of assessment that actively cultivates human potential.
The Historical Shift: From Sorting to Cultivating
The limitations of conventional testing were recognized long ago. As early as the 1950s, educational psychologist Edmund W. Gordon, working with Else Haeussermann, observed that children labeled “uneducable” thrived when given tailored support, not standardized tests. Haeussermann’s approach focused on identifying conditions for success rather than merely diagnosing deficits. This fundamental insight—that assessment should reveal potential rather than simply measure what exists—is now being revisited through the lens of AI.
From Check Engine Light to GPS Navigation
The Gordon Commission on the Future of Assessment (2013) argued that traditional standardized tests create an artificial gap between testing and teaching. Current systems act like a “check engine” light—alerting you to a problem long after it has occurred. What’s needed is a real-time “GPS dashboard” that guides learning, not just reports results. This means moving beyond simple measurement to understanding why a student struggles, and what interventions might help.
Dynamic Pedagogy: The Integrated Approach
The key is shifting from evaluating outcomes to supporting processes. Instead of simply measuring a plant to judge its health, we measure its needs (water, sunlight, soil) to help it grow. This principle translates directly to education, through strategies like dynamic pedagogy —where assessment, curriculum, and instruction work together seamlessly. AI-powered learning platforms like Khanmigo and game-based systems already demonstrate this potential, providing real-time feedback and personalized challenges.
Human Variation: A Strength, Not Noise
Traditional testing often fails to account for the richness of human differences. Factors such as cultural background, motivation, and cognitive style are treated as “noise” to be minimized, rather than assets to be leveraged. The question shouldn’t be “How smart is this learner?” but “How is this learner smart?” Embracing this diversity is crucial for personalized learning.
The Pedagogical Troika: Assessment, Teaching, and Learning
The recently released Handbook for Assessment in the Service of Learning (2025) solidifies this vision through the metaphor of a three-legged stool: Assessment, Teaching, and Learning. Removing any leg destabilizes the whole structure. Without feedback and insight (assessment), learning suffers.
The AI Imperative: Scaling Personalized Learning
While cost has historically limited personalized education, AI now enables pedagogical analytics at scale. AI can power a learning “GPS,” providing step-by-step guidance instead of a final “verdict.” This technology isn’t just desirable—it’s increasingly practical. The future of education hinges on embracing this shift.
The integration of AI into assessment isn’t merely a technological upgrade; it’s a fundamental reorientation of how we understand learning itself. By prioritizing insight over ranking, and growth over judgment, we can unlock the full potential of every student.
