In the context of ongoing digital transformation in education, the search for technological tools that foster personalized learning and deepen student engagement has become increasingly relevant. This article presents the design and evaluation of a set of custom language prompts for artificial intelligence models, aimed at enabling AI to act as an interactive, home-based tutor for high school chemistry students.

The study comprises two interconnected parts. The first outlines the theoretical foundations of the project, drawing upon principles of social constructivism, tutoring approaches, and personalized learning methodologies. It further details the design of prompts that reflect didactic communication, graduated difficulty levels, and empathetic instructional style. The prompts, organized according to core topics in chemistry—such as redox reactions, organic compounds, and states of matter—serve as interactive conversational scenarios intended to support autonomous learning and the development of critical thinking skills.

The empirical component involves a comparative study with two groups of students: an experimental group using the AI-supported prompts and a control group relying on conventional learning methods. Quantitative and qualitative data—including knowledge tests, student reflections, and teacher observations—were analyzed to assess the pedagogical effectiveness of the AI intervention.

Findings indicate that strategically designed AI interactions hold significant potential in enhancing conceptual understanding, fostering learner autonomy, and increasing motivation in science education. The results suggest practical implications for educators and researchers seeking to integrate conversational AI technologies into secondary-level STEM instruction.

Keywords: Artificial intelligence in education, language prompts, personalized learning, evaluation of teaching effectiveness, secondary.