AI Creates the Patient and the Student Does the Reasoning

What she’s doing: Kari Jensen coordinates BIOL 1108, Foundations of Biology Lab, a freshman-only course for biology majors, and teaches BIOL 1100 and 1101 for the College of Professional Studies (CPS). The lab has long included a library assignment on finding, citing, and critically summarizing scientific literature. This year, in collaboration with Missy McElligott, that assignment became an investigation of AI as a research tool, with template design support from CATLR’s Klaudja Caushi.

In Part A, student pairs choose a topic, ideally one tied to an active Northeastern lab from a provided list, and write a question about the current state of the literature. Kari models prompt specificity: a weak prompt asks what is happening with CRISPR; a stronger one names in vivo base editing for inherited blood disorders in humans and asks for primary research from the last three years. Students run the identical prompt through Scopus AI, available through the university library, and a second tool of their choice, ask each to name its sources, and complete a comparison table and bias analysis. In Part B, students track down a cited paper and reflect: Does it exist? Does the AI summarize it accurately?

What’s working and what isn’t: Kari piloted the assignment this summer with about 15 CPS students, using antibiotic resistance as a shared topic. Students enjoyed it, and the comparison unsettled loyalties: some who favored one tool preferred the other’s analysis. A pattern emerged from the reflections. Scopus AI, searching paywalled library holdings, surfaced stronger sources, while open tools often cited low-ranked or obscure journals, translated work, and opinion pieces lacking data, treated as fact.

One friction was apprehension. Even with the activity framed as an investigation of AI, students asked whether using it in class was truly allowed. Many were post-baccalaureate students who already hold degrees, and Kari noticed considerable skepticism in that group. Another student with a premium tool produced thorough but slow output and seemed reluctant to spend its capacity on the task. Deep research features were optional, and none were used; Kari sees them as a possible addition.

The activity works for two reasons. Holding the prompt constant turns differences in output into evidence about where each tool gets its information. Ending at the primary source moves students from evaluating a summary to checking a claim. Anchoring topics in campus research groups adds a second payoff: students discover the research happening around them.

What’s next: The fall rollout to roughly 150 first-year students is itself a second pilot. Kari expects younger students to bring more comfort, and her goal is that they leave able to write a precise prompt, stay skeptical of output, and verify what they receive. Fall results will shape any Part 3, and the assignment is slated to become a permanent part of CPS General Biology I and II.

Have you tried something new with AI? How did it go? Send an email to [email protected] to let us know!