Education

Educational institutions are adopting systems that make consequential judgements about children and young people — proctoring, behavioural monitoring, attainment prediction, attention tracking, automated marking — faster than they are building any capacity to evaluate them. Adoption is often driven by procurement cycles, budget windows and vendor relationships rather than by evidence of benefit, and the population affected is among the least able to object.

There is a second problem, in the opposite direction, and the two are connected. Technical education still treats ethics as a discrete subject: a single module, usually late in the degree, taught separately from the engineering and frequently by someone from another department. The syllabus is often good. The structural lesson is not, and it is the structural lesson that is retained — that ethics is what you consider after the real work is finished, by a different kind of specialist, with no bearing on whether your design is approved. The engineers who build monitoring systems for schools, and the administrators who buy them, were both taught that.

We work on both ends. On governance: what an institution should be able to demonstrate before deploying a system that judges a child, including validation in a comparable population, a route for a parent to contest an outcome, and a defined point at which the institution would stop. On curriculum: assignment material that puts the ethical question inside the technical brief — a specification that omits a population, a design that must include a decommissioning plan, a code review where the defect is a defensible-looking default with a distributional consequence. All of it released for reuse, because integration only happens when the marginal cost to a technical lecturer is close to zero.