Piloting a Machine Learning-Powered English Tutor

Research has shown that for elementary schoolchildren in India, developing literacy and fluency in English can have a large impact on university admissions, skill development, and lifetime earnings.

India faces a growing shortage of qualified English teachers and tutors that disproportionately affects poor populations.

The implementing partner (IP) developed a digital personalized adaptive learning platform for English literacy that uses machine learning and natural language processing to detect and diagnose speech errors made by elementary school students while reading aloud.

The IP then delivers research-backed, customized reading instruction ranging from word recognition skills to reading comprehension.

The platform builds on over 30-years of lab-based research and has reached over three million students in the U.S. and 17 additional countries. The IP will pilot and test their product with approximately 4,700 low-income primary school students throughout India.

The team will investigate implementation challenges, student take-up, engagement, and retention with the aim to apply what they learn in adapting the platform for scale in Indias government and affordable school sectors.

This approach has the potential to improve reading skills while also generating information on the use of machine learning to improve English literacy in low-income communities, contributing to an important body of knowledge.