You are currently viewing When AI Enters the Room: How Early Framing Shapes Student Decision-Making

When AI Enters the Room: How Early Framing Shapes Student Decision-Making

During a recent interview for a longitudinal study on undergraduate writing and learning that I am conducting at my school, one student described a course he was currently taking with a professor he admired deeply. He spoke warmly about the instructor’s passion, clarity, and care for students. He also described the professor’s stance on generative AI as “very strong” and “very clear.” AI, the student explained, was framed as a threat to learning, something to refuse rather than explore: “He says it’s making the world stupider.”

Later in the interview, when asked how much students in that particular course still used AI despite its banishment, the student paused, laughed a little, and said something like, “A lot.” He went on to describe his own use, which came across as careful, strategic, and largely unnoticeable. As the student reiterated, he respected the professor; at the same time, however, he was distinguishing between what was supporting his learning and what was not. That moment has stayed with me because it revealed something instructive: While the professor’s position was clear, it was less obvious how students were meant to reason about their own choices within the narrow bounds of that position. Rather than being an outlier in our study, we heard the sheepish “a lot” response often enough to recognize it as a pattern rather than an exception.

In many courses, generative AI enters the room long before students open a tool for one of our assignments. It appears in a syllabus statement, a first-day slide, or a brief announcement meant to establish expectations early. To instructors, these moments often feel administrative, but to students, they function as cues about the kind of learning environment they have entered. Before they even submit their first assignment, students are already interpreting what is valued and what is risky. Is uncertainty treated as part of the work or as something to avoid? Are students expected to exercise judgment or simply to comply? Are expectations framed as learning commitments or as boundaries not to cross? These early impressions shape how students think about AI as well as how visible they allow their own learning to be.

As I’ve observed in working with faculty from various types of institutions, conversations about generative AI tend to focus on enforcement: what students are allowed to do, what they must disclose, and what will happen if they cross a line. Far less attention is paid to how this early framing shapes students’ motivation to engage openly in the learning process. Research on learning motivation suggests that this omission matters. Ambrose et al. (2010) define motivation as the personal investment a learner has in reaching a desired outcome and note that motivation “generates, directs, and sustains what students do to learn” (p. 69). In their synthesis of the research, three conditions consistently support motivation: Students must believe they can succeed, must perceive value in the task, and must experience the learning environment as supportive.

Each of these conditions is sensitive to how instructors frame expectations. When students believe that missteps will be cause for punishment rather than examination, their expectation of success narrows. When students don’t perceive how policies connect to learning goals, the value of the work becomes instrumental rather than intrinsic. When uncertainty becomes a liability, students learn quickly that discretion is safer than exploration. John Dewey’s work sharpens this point further. In Democracy and Education (1930), he distinguishes value from interest and argues that interest is not a matter of preference or enjoyment but the energy that emerges when intellectual curiosity meets purposeful activity. Interest, in this sense, determines how much time and effort learners are willing to invest. When students see no safe way to test ideas, ask naïve questions, or reason aloud, interest often redirects itself toward efficiency and self-protection.

This dynamic is especially visible in conversations about AI. Students regularly report high levels of external motivation for attending college, citing factors like employment and financial security, and public discourse reinforces this framing through rankings, earnings data, and return-on-investment calculations. At the same time, recent research on student well-being suggests that many students are anxious, overextended, and keenly aware of the risks associated with academic failure. In such contexts, students tend to make pragmatic decisions. If a learning environment signals that judgment is risky and process talk is unwelcome, students are unlikely to disclose how they are actually navigating tools like AI even when they care deeply about learning.

This helps explain a paradox I encounter frequently in my study’s interviews and classrooms: that students can respect their instructors, value learning, and still choose to conceal their decision-making. The issue is not a lack of motivation; it’s a learning environment that makes visibility feel unsafe. Seen this way, the pedagogical challenge AI poses is about whether our instructional choices create conditions that encourage students to think openly, reason publicly, and practice judgment or whether they motivate students to comply efficiently and say as little as possible.

The student I mentioned at the beginning of this article was not confused about his professor’s stance on AI. He understood it clearly and even respected it. What he did not experience was an opportunity to think aloud about how that stance connected to learning goals, disciplinary practice, or his own development as a writer. In that context, students adapted. AI use didn’t cease; it became sneaky, although I cannot say in a malicious way. Students made decisions privately rather than discuss them openly, and questions about process gave way to strategic silence. As much as I, too, respect my colleagues’ decisions on banning AI, hiding one’s AI usage feels like a predictable response to instructional conditions that reward invisibility over deliberation.

My sense is that faculty might worry that explaining their reasoning around AI will weaken their authority. Yet work on self-determination and learning suggests the opposite. When faculty ground expectations in clear learning rationales rather than surveillance or threat, students are more likely to internalize them. Authority does not require suspicion to function; it requires coherence. Students are more likely to respect boundaries when they understand how those boundaries support learning rather than simply limit behavior.

One way I have tried to address this gap is by shifting where AI conversations begin. Rather than starting with what instructors will allow or prohibit, I ask them to articulate the learning priorities or values that matter most in their courses before drafting any policy language. Instructors identify one or two things they most want students to develop—such as independent analysis, collaboration, or fair assessment of individual work—and then draft a short statement explaining how their expectations around AI support those priorities. When instructors share these statements with colleagues, the differences are often striking. Two courses may prohibit AI entirely yet sound markedly different to students depending on whether the rationale centers on equity, skill development, or disciplinary norms. Similarly, courses that permit AI use vary widely in how explicitly they name the judgments students are expected to make.

This reframing does not eliminate tension around AI, nor should it. Generative AI raises real questions about authorship, effort, expertise, and a slew of other ethical issues. But when instructors make the reasoning behind their expectations explicit, students are more likely to open up about their own decision-making rather than hide it. The challenge, then, is not choosing the “right” AI policy but making visible the instructional logic behind a particular stance so that students can practice judgment rather than conceal it.

References

Ambrose, S. A., Bridges, M. W., DiPietro, M., Lovett, M. C., & Norman, M. K. (2010). How learning works: Seven research-based principles for smart teaching. Jossey-Bass.

Dewey, J. (1930). Democracy and education. Macmillan.


Kristi Girdharry is associate professor of English at Babson College. She is a cofounder of The Generator, an interdisciplinary AI lab. Her forthcoming book, Getting Learning Right: The Promise of Higher Education (coauthored with Chris W. Gallagher and Kevin G. Smith) will be published in August 2026 (MIT Press).

The post When AI Enters the Room: How Early Framing Shapes Student Decision-Making appeared first on The Teaching Professor.