A consortium of North American and European school districts began testing a privacy filter for classroom tutor logs on Sunday, aiming to separate useful learning records from personal information captured during home study sessions.

The move follows a week of questions from parents and teachers about Level-7 educational assistants, which can adapt lessons, monitor attention patterns and recommend interventions with minimal human prompting. While most districts reported smoother homework completion, several also acknowledged that raw logs included sibling conversations, meal schedules and health references irrelevant to schooling.

Global AI capability is rated at level 8 this year, but education authorities have generally limited classroom products to narrower systems with stronger oversight. The new filter would summarize student progress, flag suspected academic distress and delete incidental household data within 24 hours unless a guardian approves retention.

Teachers' unions welcomed the review but warned against treating privacy as the only concern. They said districts also need time for staff to interpret tutor recommendations, challenge faulty assessments and explain to students when a machine-generated comment will or will not affect a grade.