Google announced new school AI controls on November 3, including an administrator opt-in for K-12 access to Gemini products and Google Search. If you teach or manage school technology, that puts a familiar research tool inside the same policy conversation as the new AI tutor.
The announcement also brings teacher-led Guided Learning to Google Classroom globally. There are two decisions to assess: whether a particular guided lesson helps students, and what access a school is choosing for its accounts. A useful tutoring feature shouldn't make the second decision harder to explain.
The controls include Search
Google's policy announcement says access to Gemini, Gemini Notebook, Gemini in Classroom and Search will start off for K-12 student accounts until administrators turn it on. Google will share implementation details in the coming months before rollout; this isn't evidence that every school account changed overnight.
Google says "schools should have control over AI in their classrooms". The inclusion of Search makes that promise consequential. The Guardian reports that schools rejecting AI would lose Search access.
A school evaluating an optional tutor may also depend on web search for ordinary research. We'd want those decisions explained separately, with clear settings for different age groups. The account administrator should be able to describe the effect of a choice before teachers have to plan around it.
What teachers can assign
Guided Learning in Classroom anchors teacher-assigned activities in course material and gives teachers information about student progress. The new Search experience is on a different schedule: a pilot in selected US schools later in November, followed by further expansion.
Anchoring a lesson in class material is a sensible starting point. It gives a teacher a basis for checking whether the assistance follows what the class is learning. Then comes the harder question: after the conversation ends, can the student take the next step alone?
A record of the interaction can help a teacher see where an explanation failed, but someone still has to interpret what the student understood, connect it with their other work and decide what to revisit. The product earns its place when that information improves the next lesson.
The research measures an early learning gain
In Google's research account, more than 300 University of Tokyo students studied statistics for half a day. The group given Guided Learning alongside standard resources improved test scores by 8.1 percentage points; the standard-resource group improved by 5.3 points. That's a 2.8-point difference in gains, not evidence of a lasting improvement across K-12 education.
A separate four-week pilot in three UK secondary schools supplied teacher and student feedback. Google used that feedback to improve visibility into student activity and keep conversations anchored to uploaded curriculum. These are useful early findings with different populations and different measures.
Keep a student's learning gain separate from a model's score on a maths benchmark, such as the MIT weight-editing experiments. The school needs to know what the learner can do after receiving help.
Guided Learning gives teachers a specific lesson format to try. Before a school enables it, the adults should be able to tell parents exactly what the access choice turns on.
Cover: AI-generated editorial illustration by TMRW using Google branding. The classroom objects and opt-in card are illustrative, not a screenshot of Google's controls.
