Stanford’s classroom AI examples separate question handling from teaching the answer

Stanford’s [October 28 report](https://ed.stanford.edu/news/stanford‑instructors‑tap‑ai‑deeper‑learning) describes instructor‑built AI tools that handle different obstacles to learning: too many lecture questions to review at once, uneven preparation for a course, and students needing guidance through a difficult problem. The projects come from earlier courses and a teaching share‑out series, rather than four products newly launched this week. For an instructor considering a similar tool, the first choice is which part of the classroom process needs help. Sorting questions for a lecturer, checking assumed background knowledge and coaching a student through a task require different outputs. A fluent answer from a general chatbot does not by itself solve each of those jobs.

Amit Kaushal, a Stanford clinical associate professor, built a system that groups, ranks and summarizes questions submitted during a lecture. It identifies common themes so he can decide what needs explanation, while also allowing students to participate anonymously. [Stanford’s account](https://ed.stanford.edu/news/stanford‑instructors‑tap‑ai‑deeper‑learning) describes his aim as preserving the relationship between lecturer and learner. Here, the AI’s output is an organized view of the audience’s questions. The lecturer still chooses what to address and supplies the explanation. That makes the useful evaluation task specific: does the grouping preserve important differences between questions, and does it help the instructor notice confusion that would otherwise be missed? Those are questions to examine in a classroom trial, not results this report establishes.

The biology example operates before a lesson reaches that point. Waheeda Khalfan and colleagues used an AI pipeline to analyze course content and the relationships between foundational biology courses, which students can take in different orders. After reviewing and refining its output, the team built first‑week self‑diagnostic assessments and preparation guides. A Foundations Navigator shows the background concepts each course assumes and which courses cover them. [Source: Stanford](https://ed.stanford.edu/news/stanford‑instructors‑tap‑ai‑deeper‑learning) The distinction is between mapping the curriculum and judging an individual student. The reported pipeline identifies assumptions in course materials; the diagnostic assessments help students explore their preparation. Human subject expertise checks the map. An instructor adapting this approach would need to verify prerequisite relationships, not simply accept an AI‑generated course sequence.

STEPS, developed by Shima Salehi and chemistry lecturer Jennifer Schwartz Poehlmann, asks students questions as they work through STEM problems instead of giving the answer. In classical Japanese, Ariel Stilerman’s BungoBot likewise responds to a submitted sentence with questions that guide the student’s interpretation rather than immediately supplying a translation. [Stanford describes both approaches](https://ed.stanford.edu/news/stanford‑instructors‑tap‑ai‑deeper‑learning) as ways to support the process of working toward an answer. These examples put a different requirement on the model: its response must move the learner’s reasoning forward without taking over the task. A useful review would look at the questions it asks, when it reveals information and whether the student can explain the next step. The university report offers instructor and student accounts; it does not establish comparative learning gains through a controlled evaluation. Across the examples, the practical design choice is where responsibility stays. A lecturer answers grouped questions, subject experts curate prerequisite materials, and learners work through a tutor’s prompts. Choosing that boundary first gives a teaching team something more concrete to assess than whether an AI tool sounds knowledgeable. AI‑generated conceptual illustration of an instructor organizing questions while adult learners work through ideas. It does not depict the Stanford instructors, students or tools described in this report. Original artwork by OpenTools Team, based on [Stanford’s teaching report](https://ed.stanford.edu/news/stanford‑instructors‑tap‑ai‑deeper‑learning).