Provider: EnlightenAI
Program Description
EnlightenAI is an AI grading and feedback assistant for student writing. A teacher grades a small sample of essays, the AI learns the teacher’s standards and proposes scores and comments for the rest, and the teacher reviews them before release. Students are meant to use the feedback to revise and resubmit. In this study, AP U.S. History teachers used it for document-based, short-answer, and long essay assignments.
Findings
- Null effect: Access to EnlightenAI had no detectable effect on students’ spring writing scores or AP U.S. History exam scores.
- Nearly all students (328 of 329) logged in, and students submitted 4,846 assignments, but almost none revised their work. No student completed more than one of the 6 to 8 intended revision cycles, and two-thirds completed none.
- No differences in effect were detected by gender, race or ethnicity, prior achievement, or lunch status.
- No cost analysis was conducted. Comparison students attended Uncommon high schools without access to EnlightenAI or a similar tool.
- Because students almost never revised their work, the study tested access to EnlightenAI, not the revision-based model it was designed to support. With only nine schools, it could also detect only large effects (about 0.30 SD or more).
| Research Partner | SETA-ED |
| Study Year | 2025-26 |
| Study Location | 9 Uncommon Schools high schools in New York, New Jersey, and Massachusetts (5 EnlightenAI, 4 comparison) |
| Modality | AI Platform with Teacher Review |
| Study Design | QED (Quasi-Experimental Design) |
| Total Sample | 599 rostered students; matched samples of 386 for writing (193 EnlightenAI, 193 comparison) and 278 for the AP exam (139 EnlightenAI, 139 comparison) |
| Characteristics of Treatment Group | Writing sample (n = 193): Male: 45% Black: 84% Hispanic or Latino: 15% Other: 2% English Language Learner: 2% Students with Disabilities (IEP): 2% Economically Disadvantaged: 82% AP exam sample (n = 139): Male: 48% Black: 78% Hispanic or Latino: 21% Other: 1% English Language Learner: 3% Students with Disabilities (IEP): 1% Economically Disadvantaged: 86% |
| Subject | Writing (AP U.S. History) |
| Grades | 9, 11 |
| Average Actual Dosage | <1 revision cycle per student |
| Scheduled Dosage | 6 to 8 revision cycles per student |
For more about Accelerate’s takeaways from this study, check out the Accelerate blog.
