What the “Tutor in Your Pocket” Could Mean for Global Learners

April 24, 2025

The ability of large language models to quickly answer questions or explain complex concepts via a smartphone app, for example, has led some to refer to them as “a tutor in your pocket.”

That capability is running into another phenomenon: the widespread availability of inexpensive AI around the world. Together, these two technologies could transform global education, offering tailored learning experiences to students in remote and underserved regions and providing unprecedented access to quality educational resources.

“Education in underserved and remote regions has long been plagued by resource scarcity, whether it’s a shortage of qualified teachers, limited access to high-quality learning materials or a lack of personalized instruction,” said IEEE Member Saptarshi Ghosh. “While the internet has opened doors, it still falls short in addressing the unique needs of individual students. AI changes that.”

AI’s Potential to Fill a Persistent Gap

AI-powered learning systems are already available. Language apps that have incorporated generative AI can now offer interactive conversational practice that mimics real life.

Extending that capability to under-resourced schools, however, is still in its infancy. Pilot projects to address this challenge have shown promise.

One project detailed in IEEE Transactions on Education showed that an AI-powered tutor had an accuracy rate of about 88%. Another experiment found that students were often able to spot the mistakes large language models made when tutoring math. However, the researchers observed that students who already knew how to solve the problems correctly were much better at identifying those errors. This led them to urge caution about relying too heavily on generative AI as a teaching tool.

More Than the Transfer of Facts

The idea that generative AI and large language models could serve as a “tutor-in-your-pocket” does have some weaknesses. A big one: education is more than the transfer of facts. A student that masters a new mathematics technique by using generative AI doesn’t really learn how to apply it in the real world.

“Education consists of three key dimensions: value shaping, ability training and knowledge transfer,” said IEEE Graduate Student Member Siyuan Sun. “While AI can significantly improve fairness in knowledge delivery, it is insufficient in fostering values and abilities, which are more crucial for holistic education.”

Before the COVID-19 pandemic, many experts hoped widespread internet connectivity and the availability of online learning would help students stay on track. In many parts of the world, though, the absence of in-person learning has been blamed for declining test scores. One question is whether a reliance on artificial intelligence would replicate that scenario.

“That’s the great challenge. If we compare this to generative artificial intelligence in education, there are some parallels, but also key differences,” said IEEE Member Gabriel Gomes de Oliveira. “The similarities include over-reliance on technology and declining critical thinking skills; the differences lie in artificial intelligence being a supplement, not a replacement, and the personalization of learning alongside teacher integration.”

Go In-Depth: Teaching the ethical use of AI must go beyond technical instruction to include insights from fields like philosophy, law, social sciences and education, a new report from the IEEE Standards Association argues.