← Cases
United StatesEducation2023 – ongoing

Georgia Institute of Technology

Generative AI Jill Watson teaching assistant

Georgia Tech reported A grades for 66% of students with Jill Watson access versus 62% without access

WorkingCompany's own claimOfficial ↗

Problem

Online students needed answers grounded in their course materials. General-purpose assistants could provide answers that did not match the course.

Approach

Georgia Tech deployed a ChatGPT-based Jill Watson in its online AI course by fall 2023. It retrieved course material and checked responses against that material. A controlled experiment compared students with and without Jill Watson access; both groups could use ChatGPT.

Results

66%share receiving A grades with Jill Watson access, versus 62% without access, according to Georgia TechCompany's own claim[1]

Takeaway

The comparison tested access to a course-grounded assistant alongside general ChatGPT availability. The university described better grades as an association with access.

Categories

Industry
Education
Business function
Production and quality
Evidence
Company's own claim

Sources

  1. Georgia Tech’s Jill Watson Outperforms ChatGPT in Real Classrooms
    Georgia Tech AI2 Sep 2025Company publication

Source summaries

[1]Georgia Tech AI2 Sep 2025

Georgia Tech described its generative teaching assistant and classroom comparison. It reported A grades for 66% of students with access, compared with 62% without access.

Full details are in the original sources.

Added 1 Nov 2026 · Updated 1 Nov 2026
Open case page ↗