AI is already changing how quickly learners can generate answers, summarize content, and draft explanations. In TBL, that shift can either flatten team discussion (if students treat AI as the final authority) or deepen learning (if AI becomes something teams interrogate, challenge, and refine). The difference comes down to design.
What AI is already doing well for TBL instructors
Many educators are using AI as a practical time-saver during course and session prep. It can help generate clinical-style vignettes, draft question stems, and propose multiple-choice options—especially for lower-order knowledge checks. It can also be useful for brainstorming alternate answer choices, common misconceptions, or ways to increase realism in cases.
The key is intention. AI outputs become far more useful when the prompt is anchored to clear learning outcomes and structured expectations. Frameworks like Role–Objective–Design–Example–Sense check (RODES) help constrain the tool’s “creativity” and keep it aligned with what you actually need. That “sense check” step is especially valuable: it confirms the tool’s understanding before you invest time reviewing a full set of generated content.
The core risk: answers without reasoning
The most common challenge is straightforward: AI makes it easy for students to reach a correct answer quickly—sometimes without the discussion that helps them learn. In TBL, that can show up as teams converging too fast, skipping debate, or relying on AI-generated explanations without evaluating them.
Rather than trying to outpace AI, educators can design TBL activities where the reasoning is non-negotiable.
How educators are maintaining deep learning in TBL despite AI
Design applications with multiple plausible answers.
When more than one option is defensible, the task becomes less about “finding the right answer” and more about comparing trade-offs, weighing evidence, and articulating a rationale, all of which encourages higher-order thinking. AI can still generate a recommendation—but it can’t replace the team’s judgment unless the activity is designed to reward shortcuts.
Make it low-stakes enough to encourage exploration.
If students feel policed, AI use becomes invisible. If the goal is to practice reasoning, teams are more likely to surface AI output and evaluate it openly. Low-stakes structures can reduce “answer hunting” and increase discussion quality—especially when paired with clear expectations about justification.
Assign an “AI accessor” role inside teams.
One practical approach is to give one student a defined role: consult AI after the team has discussed the problem, bring the output back, summarize it, and invite critique. This keeps AI from becoming a private shortcut for each student. Instead, it becomes a shared input the team can test together—more like a confirmation or challenge point than the first answer. The team’s job is still to reason, question assumptions, and decide what holds up.
Use AI’s speed to deepen the task, not shorten the class.
If teams finish faster, the solution isn’t to end early—it’s to raise the level of thinking. Add a twist, introduce a second layer of context, require justification plus a critique of an alternative, or ask teams to identify what evidence would change their decision.
Responsible AI use: policies vary, but expectations can still be clear
Institutional guidelines are evolving and often discipline-specific. Even when policy is unclear, educators can set course-level guardrails that are concrete and teachable: what counts as acceptable support, what requires disclosure, and what students remain accountable for.
Transparency helps too—especially when instructors model responsible use (for example, acknowledging when AI supported drafting while emphasizing human oversight). Responsible use should also include verification. AI can produce confident but inaccurate answers, so students need clear expectations for how to check AI-generated output before relying on it.
In TBL, this can be built directly into team discussion: ask students to support their answers with evidence, compare AI suggestions against course concepts or reliable sources, and explain why their final decision is stronger than the alternatives. When AI output conflicts with what students know—or when different teams receive different AI suggestions—those moments can become useful prompts for deeper reasoning rather than simple errors to avoid.
A practical takeaway
AI is not automatically a shortcut or a solution. In TBL, it becomes useful when it pushes teams toward better questions, clearer justification, and stronger critique. If your applications demand judgment—and your facilitation makes reasoning visible—AI can raise the quality of team learning instead of eroding it.
Explore how AI is shaping TBL teaching and download our whitepaper for practical insights from our latest survey findings.
