Generative AI Pilots
The Center for Teaching Innovation is collaborating with campus partners at Cornell to pilot several GenAI bots for instructor use in the classroom. One of the goals of these pilots is to help determine whether or not – and the extent to which – these models can benefit student learning.
Each of the below learning technologies is designed to allow instructors to stay in-the-loop during student-AI interactions; is highly customizable to individual courses; offers timely, instructor-designed formative feedback to students based on course content; and creates a protected environment for students to upload their work.
Explore each pilot below:
To learn more, or to participate in one of these pilots, please fill out our form. If you're interested in piloting an online locked-down testing environment, we recommend exploring our Respondus LockDown Browser pilot.
MyEssayFeedback
MyEssayFeedback utilizes commonly available and customizable AI models to provide students with detailed and timely formative feedback on their written work, using the style, voice, and other rubric criteria designed by their instructor.
Students submit their written work within a protected environment, then engage with a conversational AI bot of their instructor’s design to receive formative feedback as they work through their drafts. The bot helps to guide students towards proficiency as established by their instructor. Meanwhile, instructors can view student–AI interactions to gain insight into the larger impact of AI on their students’ learning.
MyEssayFeedback also provides instructors with access to a curated collection of feedback templates designed and shared by peers at universities worldwide. This optional open education resource can help to jump-start the design of course-specific activities, while helping instructors articulate their own course-specific rubrics.
The Process
- Instructors design their learning activity and feedback style/criteria, either with the help of templates, or without. They can test, iterate on, and perfect their design before sharing with students.
- Students access the activity through Canvas. They upload their draft, receive AI-generated conversational feedback designed by their instructor, and can then choose to use that feedback or discuss it further with the AI bot. The process is meant to provide students with instructor-like feedback as they write, and they can submit as many drafts of their work as instructors allow.
- Instructors can view the student-AI interaction if they choose, in order to gauge whether or not the interaction was helpful to students, and/or to judge the quality of the AI feedback. An optional space for student reflection on the feedback they received and the writing process can also be added.
Why We Like It
MyEssayFeedback gives students equitable access to timely feedback as they write. Feedback is also designed by individual instructors for their specific course, which allows them oversight into the student-AI interactions. Instructors can reflect on, refine, and articulate their feedback methods through designing and testing an AI bot. This process also helps to build instructor competency in the use of AI, as they learn to customize and write effect prompts. For students, the AI model functions as a thought partner, rather than as a way to copy/paste unoriginal work. Student-AI interactions are transparent, and the system allows equitable access to an AI bot where data is stored in a secure environment.
Ed Discussion Bot++
Ed Discussion Bot++ is a chatbot deployed in an Ed Discussion course where the Bot has been trained on the course’s content. Student-AI interactions include answering student questions and providing additional guidance for student learning using course-specific materials.
This exchange provides a student-centered approach for engagements with course materials or provides students with critical information regarding class specifics in a timely manner.
Meanwhile, the process centers the instructor’s expertise by drawing from only those materials shared with the bot, rather than relying on generalized information. The instructor or TA can preview Bot answers and endorse them, or allow the bot to respond to students automatically, without their review.
The Process
- The instructor creates an Ed Discussion course where the Bot++ is activated.
- The instructor uploads documents for the bot to draw from for its responses.
- Students ask course-related questions in the Ed Discussion course site.
- The Bot++ answers questions that instructors can edit/approve as needed.
- Instructors can set limits on the length of conversation that students can have with the Bot++ (i.e., the number of follow up questions students can ask), as well as its method of answering (i.e., by guiding learning or providing answers).
Why We Like It
The Ed Discussion Bot++ system incorporates the use of well-designed and stable, licensed technology that is already used by thousands of students at Cornell. Ed Discussion’s bot offers a dynamic and student-centered way to interact with course content via discussion threads that reference materials uploaded by the teaching team. Potentially, using the Bot can reduce instructor/TA time spent on course management by automating responses to commonly-asked student questions.
HiTA.ai
HiTA is a conversational AI tool designed to provide students with course-specific activities by drawing only from instructor-provided content. The system emphasizes instructor oversight and pedagogical alignment, providing students with timely hints, scaffolding, or answers that support learning, while allowing instructors maintain control of the process.
HiTA enables students to interact with a bot trained on course readings, lectures, and assignments. The tool’s design encourages independent problem solving by emphasizing guidance and critical thinking, rather than simply delivering answers. Instructors can monitor student-AI interactions, refine the bot’s behavior, and gain insights into common misconceptions or gaps in understanding.
HiTA can be configured to support modules, lessons, activities, assignments, and grading, as well as support personal assistant bots for instructors.
The Process
- Instructors upload course materials (e.g. readings, assignments, lecture notes) into the HiTA system.
- The instructor configures the conversational style, tone, and level of scaffolding — for example, whether the bot should give direct answers, hints, or ask guiding questions.
- Students interact with HiTA to ask questions, check understanding, and receive feedback or explanations based on course materials.
- Instructors have the option to preview, approve, or edit bot responses, or allow the bot to respond automatically.
- Student–AI interactions can be reviewed by the instructor to assess helpfulness, identify patterns, or better understand student needs.
Why We Like It
HiTA keeps instructors central by grounding AI responses in course-specific materials, ensuring alignment with teaching goals. It offers students equitable, always-available support that encourages deeper engagement with their coursework, while reducing repetitive questions for instructors and TAs. Student–AI interactions remain transparent, giving faculty insight into how the tool shapes learning.