AI in Education Needs a Learner-Rights Compact

A learner can be required to use a platform without having chosen the artificial-intelligence features embedded inside it. A teacher can be told to innovate without being given evidence, training or authority to switch the feature off. A parent can receive a general privacy notice without learning whether a child's prompts, voice, behaviour, disability information or academic struggles will be analysed, retained or used to shape future decisions.

These are not marginal implementation problems. They reveal a governance gap between institutional adoption and human permission. Buying, licensing or activating an AI service does not by itself establish that the use is educationally justified, legally proportionate or acceptable to the people whose learning and data make the system possible.

Education therefore needs more than an AI policy. It needs a learner-rights compact: a public and enforceable account of what the technology is for, what it may collect, what it may decide, what people can refuse, what must remain human, how competence will still be demonstrated and who can stop the system when evidence or safeguards fail.

On 21 October 2026, United States Senators Richard Blumenthal and Josh Hawley opened a bipartisan inquiry into Google's promotion and default integration of AI products in schools. Their letter asks why some AI functions cannot be disabled, what research the company has conducted on educational and emotional effects, what concerns…