Teaching judgment in the age of answers
Why education should focus less on answers and more on calibration, criteria, and courage to choose.
Education has long used answers as evidence of understanding. Generative AI weakens that proxy: a learner can now produce a fluent answer before developing the judgment required to evaluate it.
What remains difficult
The hard work moves toward recognizing weak evidence, selecting appropriate criteria, testing an explanation, and deciding when confidence is justified.
These skills are less visible than a completed response. They require students to show how their view changed, which alternatives they rejected, and what would cause them to reconsider.
Designing for calibration
An AI-assisted classroom can ask students to predict before generating, critique before accepting, and compare several plausible answers against a shared rubric.
The goal is not to keep AI outside the classroom. It is to make judgment—not output—the central object of learning.