Google's Gemini AI Enhances Math Tutoring and Classroom Engagement in UK and Italy Studies

Google's DeepMind and Google for Education recently published studies evaluating the impact of their Gemini AI in classrooms across the United Kingdom and Italy. The research focused on how Gemini-powered tools support students and teachers in real educational settings, aiming to assess both learning outcomes and classroom engagement.

In the UK, a 10-week randomized controlled trial with over 1,100 students aged 12-15 examined Gemini's AI tutoring within Eedi's math platform. The AI tutor combined Gemini's capabilities with Eedi's pedagogical resources to provide tailored, step-by-step math support when students answered incorrectly. Human experts reviewed each AI response before delivery. The results showed that AI tutoring matched human tutors in effectiveness for many students, promoting active problem-solving rather than simply giving answers.

Despite these positive outcomes in daily assignments and quizzes, the AI tutoring did not yield statistically significant gains on the independent STAR standardized math test during the trial. Researchers attribute this to AI sessions requiring more time per problem, resulting in fewer total exercises completed and hence less curriculum coverage over the trial duration, suggesting longer-term studies are needed to assess sustained impact.

In Italy, a yearlong study with 1,600 teachers and over 24,500 students across 31 schools explored how educators used Gemini to co-design lessons. Teachers reported increased creativity and transformed classroom dynamics, with traditional lectures decreasing by 38% and student-led collaborative learning nearly tripling. Student engagement rose, as did peer discussion and ownership of learning.

Importantly, using Gemini for lesson design significantly improved teacher wellbeing by reducing the effort needed to maintain student engagement. AI-assisted lessons more often generated intrinsic activity-driven momentum, alleviating the daily stress on educators to keep classrooms energized.

These studies underscore Gemini's potential as an AI tool that supports rather than replaces educators, enhancing personalized learning and fostering active classroom environments. However, limitations include the need for human oversight in AI responses and the current focus on short-to-medium term evaluations. Future research aims to extend trial durations and explore broader educational outcomes.

Overall, Gemini demonstrates promise in augmenting math education and lesson planning, highlighting that effective AI integration in schools involves collaboration between teachers and technology to optimize learning and wellbeing.