Every AI Education Pilot Needs a Public Learning Protocol

A school announces an artificial-intelligence pilot. The platform promises personalised learning, leaders promise innovation and the first cohort begins using the system. Because the word pilot sounds cautious, the institution assumes that responsible experimentation is already taking place.

Yet no public document explains the learning claim, why these learners were selected, what comparison will be used, which teacher practices must remain, what data the provider receives or what evidence would cause the institution to stop. If results appear favourable, expansion is anticipated. If difficulties arise, they may be described as implementation problems rather than evidence about the model.

That is not yet a learning pilot. It is an early deployment with an uncertain evaluation.

Every AI education pilot needs a public learning protocol: a proportionate account of purpose, participants, comparison, human support, protection, cost, evidence, unequal effects, decision ownership and the conditions for scaling, redesigning or stopping.

Houston Independent School District announced on 31 July that two elementary schools will use personalised learning modules developed by Alpha School when the school year begins on 10 August. According to the district statement reported by the Houston Chronicle, learners who achieve high scores on district reading and mathematics assessments will spend part of the morning using the modules before rejoining classmates.

The district says the pilot will not replace teachers, carries no cost to the district and…