Google's Gemini AI Enhances Math Learning and Teaching Engagement in UK and Italy Schools

Image: Google · Source

New research shows Gemini AI tutoring matches human tutors in UK math classrooms and aids Italian teachers in creating interactive lessons that boost student participation and reduce teacher stress.

Google has released new findings on the impact of its Gemini AI in educational settings across the United Kingdom and Italy, demonstrating significant benefits for both students and teachers. The studies highlight Gemini’s role in providing effective personalized math tutoring and enabling educators to foster more interactive and engaging classrooms.

In the UK, a rigorous 10-week randomized controlled trial involving over 1,100 students aged 12 to 15 evaluated Gemini’s capabilities through a collaboration with the learning platform Eedi. When students answered math questions incorrectly, they received either static hints, help from a human tutor, or interactive AI tutoring powered by Gemini combined with Eedi's pedagogical tools. Importantly, experts reviewed AI responses in real time to ensure quality and appropriateness.

The trial revealed that AI tutoring using Gemini supported students’ math learning on par with human tutors across various assignments and quizzes. Notably, Gemini emphasized guiding students through problem-solving processes rather than providing direct answers in nearly 99% of sessions, encouraging active thinking and self-correction. Despite improved understanding and pacing, the AI-assisted students tackled fewer questions per session, which meant that short-term standardized test improvements were not observed. Google plans follow-up studies over a full academic year to further explore the long-term learning impacts.

In Italy, a yearlong study involving 1,600 teachers and over 24,500 students across 31 schools investigated how Gemini could enhance teaching practices and classroom dynamics. Teachers co-designed lessons with Gemini, which led to a substantial reduction in traditional lectures by 38%, replaced by workshops, group activities, and peer discussions that nearly tripled student-led collaborative learning. Student engagement rose from 77% to 84%, with more active participation and ownership of learning.

Additionally, teachers’ wellbeing improved considerably. The dependence on their personal energy to maintain student engagement dropped by 66%, while lessons where the activities themselves sustained momentum more than doubled. Educators reported feeling more creative and better supported, even on days with limited preparation time, thanks to AI assistance in adapting materials.

These findings underscore that AI tools like Gemini can effectively complement human educators rather than replace them, promoting active learning and reducing teacher stress. Google emphasizes that the best educational outcomes occur when technology empowers teachers and stimulates deeper student engagement.

More detailed technical and implementation reports are available for educators and researchers interested in the studies. Google’s work with Gemini represents a significant step toward integrating AI responsibly and effectively within classrooms to improve learning and teaching experiences.