Designing responsible AI for learning

Google is rolling out Guided Learning features in Google Search and Google Classroom to help students develop critical thinking by guiding them through complex problems step-by-step instead of giving direct answers. The tools act as a personal coach, using open-ended questions and structured breakdowns to encourage productive struggle and deeper understanding. Teachers can assign these activities in Classroom, where students receive tailored support aligned with classroom material while educators gain real-time insights into class progress. The feature was tested with educators and researchers to ensure alignment with classroom needs before its global launch.

Guided Learning is now available in Google Classroom worldwide, where it functions as a one-on-one study tool for students tackling assignments. Teachers can monitor student progress and identify areas needing additional support through real-time data provided within the platform. The tool is designed to keep students focused on learning objectives while fostering independent problem-solving skills. Google collaborated with schools and educators during development to refine the feature based on practical classroom feedback.

Google is also updating its approach to school controls for AI tools, including Gemini, Gemini Notebook, and Search, to give administrators full discretion over access for K-12 student accounts. Access to these tools will default to off for students until school leaders explicitly enable them, ensuring alignment with educational goals. The changes follow consultations with schools, teachers, researchers, and policymakers to address concerns about responsible AI use in classrooms. Schools will receive detailed guidance in the coming months to prepare for the transition.

The company emphasizes its commitment to responsible AI in education, aiming to support students and educators while addressing concerns about shortcuts and learning outcomes. Google states that Guided Learning has demonstrated measurable learning gains in research and is expanding its availability based on positive feedback. The updates reflect a broader effort to balance innovation with control, ensuring technology serves educational needs without compromising learning integrity. Administrators will have tools to tailor AI access to their students' needs.