AI study: 42% engagement boost for disabled students in rural New Mexico

A study published on arXiv (2606.00034) by researchers Uloma Egondu Nelson and Gil Gallegos demonstrates how artificial intelligence can dramatically improve science education for disabled students in rural New Mexico. The study employed a mixed-method design combining multiple linear regression and an Artificial Neural Network (ANN) model. It involved 120 students in grades 6 through 10 and 15 instructors across four rural schools. The AI-based learning intervention predicted student performance with high accuracy (R² = 0.92, p < 0.05). Experimental results showed a 32% improvement in science concept retention, a 27% increase in laboratory performance, and a 42% rise in student engagement after the intervention.

These findings suggest that AI-driven pedagogy can serve as a powerful equalizer for disabled learners in underserved settings. The study concludes that AI offers a promising path toward equitable science education, particularly where resources are scarce. By personalizing instruction and providing real-time feedback, the system helped bridge gaps that traditional methods could not address. The paper highlights how technology can unlock potential for students who have historically been left behind.