Skip to content
Re:Porter
WritingResourcesVisualAboutNow

Re:Porter

Notes about education, technology, and society — written from somewhere quieter than the classroom.

AIAI & EducationPAPedagogy & AssessmentTLThinking & LearningTTTechnology & ToolsCPCreative PracticeZOZoom Out
WritingResourcesVisualAboutNow

Words and work by Sam.

RSS
←Visual essays

Visual essay

Small Gains, Clearer Boundaries

A scrolly editorial note on where AI genuinely helps in schools, where it only sounds helpful, and how a responsible workflow keeps judgement in the open.

AIAI & Education·12 April 2026

Reading lens

Small gains. Higher-stakes boundaries. Clearer workflows.

This is not an argument for magical automation. It is an argument for locating the real gains precisely enough that schools can keep responsibility visible.

Most school conversations about AI still swing between two bad poles: either it changes everything overnight, or it is dismissed as a passing trick. Neither framing is useful.

What matters in practice is narrower. These tools help most where the cost of a bad first pass is low, the need for speed is high, and a human is still close enough to reshape the output before it matters.

That sounds less dramatic than the hype cycle. It is also closer to how the work actually feels.

Section 1

Three kinds of lift

Planning

Fast first passes

Useful when the blank page is the bottleneck, not the judgement.

Adaptation

Quicker revision

Rephrase, scaffold, trim, translate, and reshape without starting over.

Administration

Small friction removed

Summaries, drafts, and routine communication where speed matters more than originality.

Step 1

It helps most at the edges of a task.

The strongest use cases are usually not the glamorous ones. They live at the edges: first drafts, rough outlines, and background summaries that help a teacher move faster toward the real decision.

That is different from saying the model did the work. It more often shortened the distance to the work.

Step 2

Adaptation is often more useful than generation.

Teachers rarely need infinite new text. They need one existing thing turned into three versions: simpler, clearer, shorter, or more scaffolded.

That is where AI often earns its keep. It reduces repetition while keeping the teacher in charge of the standard.

Step 3

The quiet administrative gains are real.

Routine email drafts, summary notes, and housekeeping language are not where the moral drama sits, but they do consume time.

When AI removes small friction from those tasks, the gain is not theoretical. It is attention returned to work that matters more.

Section 2

Where judgement stays human

Routine communication

Low stakes, repeatable, usually checkable.

Lesson sequencing

Grading and feedback

Pastoral judgement

Easy to automateHarder to automate
Low consequenceHigher consequence

Step 1

Start with the low-consequence work.

If a task is routine, easily reviewed, and low consequence when wrong, AI can be worth trying. That does not mean blind trust. It means the cost of checking is still manageable.

Step 2

Planning still needs a teacher’s map of the terrain.

Lesson sequences live in the middle. Models can suggest patterns, examples, or alternative framings, but they do not know your students, your timing, or the real trade-offs in the room.

Step 3

Assessment and pastoral work sit much higher on the risk curve.

The closer a task gets to judgement, consequence, and trust, the less persuasive the speed argument becomes.

A system that sounds fluent is not the same thing as a system that is accountable.

Step 4

Some decisions should remain obviously human.

There is value in drawing a line teachers and students can understand. The point is not technical possibility. The point is where a school wants responsibility to remain visible.

Section 3

What a useful workflow looks like

1

Draft

Use the model to create a first pass quickly.

2

Check

Interrogate claims, tone, bias, and missing context.

3

Adapt

Reshape the output to fit the class, task, and standard.

4

Attribute

Be clear about what the tool did and where judgement stayed human.

Step 1

Use it to move, not to finish.

The first step is a speed step. You use the tool to get a draft on the table quickly enough that real editorial work can begin.

Step 2

Checking is not optional admin.

This is where the professional work sits. Claims need checking. Examples need testing. Tone needs calibration. The machine does not know which error will matter most in your setting.

Step 3

Adaptation is where expertise becomes visible.

The same draft looks different once it is rewritten for Year 9, for staff, or for parents. That translation is not clerical. It is the work.

Step 4

Clarity about authorship keeps trust intact.

Schools do not need mystical language about co-creation. They need plain statements about what the tool did, what the human checked, and where responsibility remained.

Closing note

The practical question is not whether AI is allowed into school work. It is already here. The better question is where it genuinely lifts the work, where it risks blurring responsibility, and where schools need firmer norms than they currently have.

That answer will never be a single policy sentence. It will be a workflow question, a judgement question, and a trust question, repeated every time a new task shows up.

Re:Porter

Notes on education, technology, and the spaces where they collide. Written by Sam Porter.

AIAI & EducationPAPedagogy & AssessmentTLThinking & LearningTTTechnology & ToolsCPCreative PracticeZOZoom Out
WritingResourcesVisualAboutNow

Words and work by Sam.

RSS