Source: Beth M. Schwartz and Christopher Hakala, Inside Higher Ed
Summary:
Beth M. Schwartz of the American Psychological Association and Christopher Hakala of Springfield College argue that while research on AI’s benefits for learning is thin, the science of how people learn is well established and can guide AI decisions now. Learning requires targeted cognitive effort: retrieving information, organizing ideas, explaining reasoning, and correcting misconceptions. Their test for any AI use is whether it strengthens those processes or replaces them. Faculty should start from the learning goal, not the tool, and institutions should fund small studies of specific AI uses, including whether students can apply what they learned without AI.
Observations:
- The authors write about college students in general, but their test applies directly to graduate business courses, where AI can now draft a case analysis, build a financial model, or write a strategy memo for the student.
- The key question is not whether an AI product “works,” but whether a specific use strengthens the thinking a course is meant to build or replaces it.
- Their productive examples keep the first attempt with the student: students explain or draft on their own, then use AI to challenge their reasoning, and decide which feedback to accept and why.
- Their call for small, iterative studies, including whether students can transfer learning to new situations without AI, fits naturally with assurance of learning.
Actions:
- For each core course, name the thinking that must stay with the student, such as framing the problem in a case, choosing assumptions in a model, or making the recommendation, and set the course’s AI policy around it.
- Redesign one assignment per course on the authors’ pattern: students produce their own analysis first, use AI to stress-test it, and submit a short memo explaining which AI feedback they accepted or rejected and why.
- Build an AI-free demonstration into assessment, such as an in-class case discussion, an oral defense, or a timed problem, so students show they can do the work on their own.
- Use your assurance of learning process to run small before-and-after comparisons of specific AI uses, and share the results with faculty.
Topic Sentence:
Until AI-specific research catches up, faculty can rely on the science of learning by asking whether each AI use strengthens or replaces the thinking students need to practice.
Citation:
Schwartz, Beth M., and Christopher Hakala. “Research on AI and Learning May Be New. The Science of Learning Is Not.” Inside Higher Ed, October 22, 2026. https://www.insidehighered.com/opinion/views/2026/10/22/science-learning-should-guide-ai-use-decisions-opinion.
0 Comments