Update on the Initiative for Pedagogical Uses of Artificial Intelligence

This is an update on an earlier post which announced a donor-funded initiative to incentivize use of artificial intelligence tools in Georgetown classrooms.

A request to the Georgetown community for proposals of innovative uses of AI yielded about 100 project ideas. Of those, about 40 were funded and are now in progress. It was interesting that the largest set of proposals came from the humanities faculty, with the sciences and social sciences tied for second place. Some projects were single investigator awards, labeled “IPAI Fellows;” others were groups labeled IAPI Cohorts. Some of the grants were awarded to students; most to faculty.

One useful classification system for the proposals gives a notion of where the innovation is targeted.

The first theme might be called, “Rethinking Ways of Teaching.” The funded projects display a wide variety of teaching innovation pilots. Some are examining using Large Language Models (LLMs) to display behaviors that might be harmful and then asking students to develop policy approaches to address those harms. Others approach similar issues from a purely ethical perspective. Still other work is using LLMs to generate text relevant to an assignment, assemble such material over different students and then have students attempt to identify which text was written by the LLM and which is of 100% human origin.

The second theme could be considered “Researching New Ways of Working.” Some work in this area is addressing whether AI might be used as an assistant to either instructors or students. Some AI-assistance might act as a tutor to students in the class. Others are examining whether AI might assist students in feedback on written exercises. Still others have students to first seek ideas from Chat GPT for issues tackled in the class, then develop or critique those ideas in their own work. There are several ideas of redesigning work to use LLMs as a starting point of a task, followed up with human completion. One project is examining how LLMs might be used as assistants to graders of exercises submitted by students. Another project uses plug-ins to Chat GPT to critique software written by students. A project uses LLMs to summarize scientific articles and then have the students critique the summary.

The third theme might be labeled efforts to improve the student experience as they navigate the university. In this group, some projects are using LLMs to write memoranda in a given style to seek decisions form administrators. Others are attempting to design a use case to help students scope out their curricular path through Georgetown. Still others are designing an AI tool to act as a career advisor to students or helping students navigate course registration each semester.

All of these ideas are consistent with an early decision that the Georgetown community made – to train future leaders, Georgetown must give them the capacity to use all the tools in existence to achieve the goals of people for others. Education at Georgetown must integrate new technologies in ways that achieve our centuries-old goals of formation and intellectual development of our students. AI and LLMs are just one among many tools that will be part of our students lives.

We are planning a second call for proposals for using AI in education and research at Georgetown, targeting a Mid-March 2024 announcement. We also plan to launch a Design Lab with activities from January to April to help prepare faculty to apply for the second round of grants, and to build the community of grantees in sharing with each other and broader faculty, staff and students. Activities to include design sessions, invited speakers, and panel discussions. All these efforts are led by Eddie Maloney of CNDLS and Randy Bass of the Red House.