Can Artificial Intelligence Expand Access to Academic Support?

Educational inequality does not only depend on the schools students can access, but also the academic support they can gain beyond school. Students from wealthier backgrounds might be able to access private tutoring, exam preparation, and other kinds of individualized support, while students from other backgrounds have fewer chances to receive additional support. As artificial intelligence is becoming increasingly common in education, an important question has emerged: can AI tutors make personalized academic support more widely available?

One main advantage of AI tutoring is accessibility. Traditional private tutoring has many limitations when accessing such as cost, time, or mobility. AI tutors, however, can provide explanations, practice questions, and feedback 24/7. They can also adjust their responses according to a student’s level of understanding. For students that cannot access private tutoring: this can be a new form of individualized academic support.

Growing evidence suggests that AI tutors can intensify student learning. In a 2025 randomized controlled study involving 194 Harvard students enrolled in an introductory physics course, researchers compared AI given instruction with traditional classroom active learning. Students who received AI tutoring achieved significantly higher post test performance across both lower- and higher-performing groups, with the difference appearing significant. These findings suggest that AI tutoring can provide effective academic support for students with different levels of prior knowledge and scientific understanding.

This study of university students suggests that well-designed AI tutors can support actual learning rather than simply providing answers.

However, AI tutoring does not automatically create greater educational equality. Its benefits depend on whether students have access to reliable internet, digital devices, and AI tools. Students in well-resourced schools may also receive more guidance on how to use AI effectively, while students in disadvantaged communities may have less access to both the technology and the support needed to use it.

The effective use of AI is also one major concern. If students directly copy answers five by the AI tutor it weakens learning rather than strengthening it. Effective AI tutoring should encourage students to come up with their own solution by asking questions, and explaining mistakes instead of immediately providing solutions.

Ultimately, tutors are unlikely to eliminate the education gap on their own. Problems such as poverty, unequal school funding, and teacher shortages are much larger and more complex. Nevertheless, AI could still reduce one important source of inequality: unequal access to individualized academic support. Therefore the real challenge is not only to make AI more effective, but also to ensure that their benefits are accessible to students from different backgrounds.

Mingyu Guan