Public teaching · The educator
You are asked to model judgment under the conditions that make it hardest to build.
For educators, programme leaders and teaching-and-learning teams working out what assessment, learning design and professional formation should become now that the artefact is cheap to produce.
01 · The double bind
Two demands that pull against each other.
Educators are asked to prepare students for work that will be done alongside AI, while assessing them on artefacts that AI now produces competently. At the same time their own conditions for developing judgment are being compressed: time, review, considered feedback.
Most responses to this have been detection-led. Detection is a policing strategy, and policing strategies do not build capability in the person being policed or in the person doing the policing.
02 · Assessment
What assessment is actually for, now that the artefact is cheap.
If an assessment measures whether a student can produce a text, and a model can produce that text, the assessment has stopped measuring what it claims. The question is not how to stop the model. It is what the assessment was really trying to detect, and whether the artefact was ever the point.
- What judgment is this task supposed to reveal?
- At what point in the process would that judgment be visible?
- Which parts of this work are legitimately delegable in professional practice?
- What would a defensible account of the student’s reasoning look like?
03 · Learning design
Designing the struggle back in, deliberately.
Learning design in this environment is largely a question of which difficulty to preserve. Not all difficulty is productive, and removing the unproductive kind is exactly what good design has always done. The new problem is that the productive difficulty is now just as easy to remove.
A worked instance: a literature review used to build the skill of tracing an argument through conflicting sources. Generated whole, it demonstrates nothing. Reframed, the student specifies what a defensible review must account for, the machine drafts, and the student is assessed on what they caught, corrected and rejected. The difficulty moves; it does not disappear.
04 · Professional formation
Forming a professional, not just teaching a subject.
Formation is the part that rarely appears in a module descriptor: what the student comes to believe they are for, what they will not delegate, and what they will escalate. It is transmitted mostly by example, which is why the erosion of the educator’s own conditions matters so much.
05 · Retained judgment
What an educator should not hand over.
- Deciding what a piece of work is genuinely evidence of
- Deciding what difficulty to preserve and what to remove
- Judging a borderline case against a standard, in context
- Deciding when a concern must be escalated rather than absorbed
06 · A bounded example
What this looks like in an assignment.
Consider a written assignment intended to demonstrate analytical reasoning. Reframed as an AI-native workflow, the human steps are specifying what a sound analysis must account for, verifying the reasoning against the evidence, and escalating where the analysis touches something the student is not positioned to judge. The drafting can move; the specification and verification cannot.
Educational illustration. Nothing here assesses the reader.
07 · Evidence status
How far the evidence goes.
This is an argument informed by learning theory and a peer-reviewed publication record in adjacent areas of online learning. It is not a report of measured student or institutional outcomes.
08 · Where this goes
Building it into a programme is Edney Learn work.
Educator workflows
In developmentEdney Learn is developing role-specific workflows for education. Nothing is bookable yet.
The public teaching stays here, free to read and argue with. Building it into a delivered programme is a different job, and that is Edney Learn’s.
Read the wider argument.
This page is one application of a larger claim about how capability is built.