Michigan U Leverages AI for Efficient Essay Grading

The university implemented an AI system to automate the grading of written assignments. The AI analyzes essays, provides initial feedback, and identifies areas where students struggle.

Client:

The University of Michigan is one of the oldest public universities in the United States. It is renowned for its academic excellence, extensive research, and strong alumni network. The university attracts students worldwide and consistently ranks among the top universities globally.

Problem Statement:

● Overwhelming workload: Large STEM courses with hundreds of essays made manual grading impractical.
● Delayed feedback: Students required timely feedback for effective learning, but manual grading caused delays.
● Identifying struggling students: Early detection of students needing extra help was crucial but difficult with manual methods.

Results:

☑️ Rapid feedback: Students receive timely, detailed feedback.
☑️ Enhanced student engagement: Automated tools boost student participation.
☑️ Early intervention: AI flags struggling students.
☑️ Faculty efficiency: Professors focus on teaching, not grading.
☑️ Data-driven improvement: Performance data informs better teaching.

AI Solution:

The University of Michigan developed an Automated Text Analysis (ATA) system to address the challenge of providing timely feedback on student writing assignments in large STEM courses.

This AI solution was designed to analyze student essays, identify key elements, and provide initial feedback.

The university collected a substantial dataset of student writing samples to implement an AI solution. Using these samples, they trained the ATA system to recognize patterns in successful and unsuccessful essays. To ensure relevance to the academic context, the system was specifically tailored to the content and structure of writing assignments in STEM courses.

Researchers from the Michigan AI Laboratory integrated the ATA system into the M-Write program, a broader initiative to improve student writing skills. This innovative approach combines automated analysis with human expertise to enhance student learning.

References:

1. How U of Michigan Built Automated Essay-Scoring Software to Fill ‘Feedback Gap’ for Student Writing

2. M-Write expands to include computer analysis in grading student essays

Industry: EdTech
Vendor: Michigan AI Laboratory

Client: University of Michigan