#  Tutor Bots and AI Feedback 

 



 ##  

  expand\_more  

 
  

 

Faculty interested in helping students learn with AI can work with us to design assistants for their courses and teach students to use AI productively. A course-specific chatbot can draw on course materials as context and follow instructions that guide its responses toward the teaching team’s goals. The aim is to offer support better suited to the course than students would get from a general-purpose chatbot alone. Alongside building these systems, we work with faculty to help students use generative AI in ways that support their learning. This can include students learning to build their own learning apps, as they have done in language courses.





 

##  A film tutor trained on course materials (GENED 1145) 

Students in GENED 1145: Global Japanese Cinema, taught by Professor Alexander Zahlten, have the option of using an AI film tutor built for the course: they ask it to define cinematic concepts and clarify key terms, which in turn supports building their film analysis skills. The tutor runs on Google Gemini and draws its answers only from a curated set of course materials, which the teaching team refined over the semester. Bok Graduate Fellow Alina Scotti, also the course's Teaching Fellow, built the tutor with the Learning Lab. A tutor like this needs one thing, a stable body of readings or lectures to draw on, and it gives students a place to ask basic questions as often as they need.

To build an AI tutor that draws only on materials you choose, contact the Bok Center.





 

##  AI-assisted pitch development (CE 11) 

In CE 11: StudioLab on Creativity and Entrepreneurship, taught by Dr. Lumas Helaire, each student develops a venture of their own. At a Learning Lab workshop on AI-assisted pitch development, students learned to use Claude to test their pitch and then to communicate it in a website. Each student organized the materials that define their venture into a context pack: a problem statement, value proposition, customer segments, and business model canvas. The pack made Claude’s feedback specific to each venture instead of generic advice, and students used it to stress-test assumptions and sharpen the pitch. Each student then built a website with Claude that presents the venture and hosts the three-minute pitch video they had recorded, with no coding background assumed. Stating the problem, the value proposition, and the assumptions clearly enough to build with strengthens the venture itself.

To design an AI-assisted pitch development workshop for your course, contact the Bok Center.





 

##  A writing companion for French learners (FRENCH 11) 

In collaboration with Wonda, Professor Nicole Mills developed an AI writing tutor and thought partner—Lee—designed for novice French writers grounded in a genre-based pedagogical approach to second-language writing. The AI writing companion scaffolds a three-stage writing process: (1) discussion and analysis of a model text’s content, style, and opportunities for textual borrowing; (2) oral brainstorming and curated vocabulary support; and (3) grammar feedback, explanations, and error identification in early drafts. Student–AI interactions are tracked in a unified platform for optional instructor review. In Spring 2024, students’ perceptions of the platform were evaluated through constructs from the psychology of language learning—writing self-efficacy, anxiety, self-regulation—and user-experience measures such as anthropomorphism, teaching presence, and enjoyment. Findings showed that students’ enjoyment and perceptions of the AI tutor were closely tied to how competent they felt as writers. Building on this work, The Learning Lab’s postdoctoral fellows are supporting follow-up longitudinal cross-sectional, mixed-methods research from Spring 2026 to Spring 2027 comparing AI-supported compositions with paper-based process-writing tasks. Pre- and post-task surveys measure students’ attitudes and perceptions, including engagement, writing self-efficacy, anxiety, perceived and affective learning, and value alignment with AI. Survey data, interaction logs, and composition scores are being analyzed to examine how AI-mediated versus paper-based scaffolding influences learner confidence, motivation, attitudes toward AI, and learning outcomes.

To measure what an AI tool actually does in your course, contact the Bok Center.