L1
·
Quiz
·
Lab
L2
·
Quiz
·
Lab
L3
·
Quiz
·
Lab
L4
·
Quiz
·
Lab
Module Test
Module 5 · Lesson 1

The Editorial Eye: What AI Actually Does When It Reads Your Draft

Understanding the difference between spell-check, grammar correction, and genuine editorial feedback — and what AI brings to each layer.
What separates a proofreader from an editor — and where does AI fall on that spectrum?

In 2022, reporter Will Douglas Heaven at MIT Technology Review submitted a piece about large language models to an internal review workflow that used an early GPT-4 preview for structural feedback. The system flagged that his third section buried the most newsworthy claim — a genuine editorial observation that a human sub-editor had missed on first pass. The story ran with that section moved to the lede. Heaven later wrote about the experience, noting the AI had caught a structural problem, not a grammatical one.

Three Layers of Revision

Most writers think of editing as one thing. Professionals divide it into at least three distinct operations, each requiring a different cognitive lens.

Proofreading operates at the surface: spelling, punctuation, typographical errors, consistent hyphenation. Rule-based. Deterministic. This is where spell-checkers have worked since the 1970s.

Copy-editing goes one level deeper: grammar, sentence clarity, word choice, factual consistency, style-guide compliance. Still largely rule-governed, but judgment enters.

Structural (developmental) editing is the deepest layer: Does the argument hold? Is the pacing right? Does the reader know what they need to know when they need to know it? This requires holding the entire work in mind simultaneously — something humans find hard and AI finds… different.

What AI Uniquely Contributes

AI language models are genuinely strong at copy-editing. They have been trained on enormous corpora of edited prose and can flag passive-voice overuse, nominalisation, hedging language, and inconsistent tense with speed no human matches at scale.

Where AI surprises most writers is at the structural layer — not because it reasons like a developmental editor, but because it can hold a 5,000-word draft in a single context window and compare paragraph 12 against paragraph 2 without fatigue or the anchoring bias humans develop after reading a draft three times.

In 2023, the literary agency Janklow & Nesbit began experimenting with Claude for preliminary manuscript notes. Agents reported that the tool consistently identified chapters where a character's motivation was inconsistently established — catches that previously required a second human reader.

Important Limit

AI cannot evaluate whether a piece is true in the journalistic sense, whether a scene is emotionally authentic to a real human experience, or whether a stylistic choice is daring versus merely strange. Those remain human editorial judgments.

The Four Editorial Modes to Know

Diagnostic Mode

You give AI the draft and ask: "What are the three biggest problems with this piece?" No fixes yet — just diagnosis. Forces the model into a structural frame rather than a line-edit frame.

Prescriptive Mode

You give AI specific criteria — "flag every sentence over 30 words," "identify every claim without evidence," "find where the reader loses the thread" — and it audits against those rules.

Comparative Mode

You give AI two versions of a passage and ask which better serves a stated goal. Useful when you're caught between revisions and need a reader reaction, not a rule.

Generative Mode

You give AI a weak passage and ask it to demonstrate an alternative — not to replace your prose, but to show you what a different approach might look like so you can decide.

Practitioner Note

The journalist and author Steven Johnson, writing in The New York Times Magazine in 2023, described using AI in diagnostic mode on his book drafts: "I stopped asking it to fix things. I started asking it to tell me what was broken. That's a completely different conversation."

Context windowThe maximum amount of text an AI model can "hold in mind" at once during a single exchange. Longer context windows allow whole-document structural feedback.
Developmental editingThe highest-order editorial work: structure, argument, pacing, character coherence, thesis clarity — distinct from grammar or style.
Anchoring biasThe human tendency to over-weight the first version of a text encountered, making it hard to see problems after repeated readings.

Lesson 1 Quiz

The Editorial Eye — 4 questions
1. Which editing layer involves evaluating whether an argument's structure serves the reader?
Correct. Developmental editing is the deepest layer — concerned with structure, pacing, and whether the reader has what they need when they need it.
Not quite. Proofreading and copy-editing work at the surface and sentence levels. Structural problems belong to developmental editing.
2. The MIT Technology Review example is significant because the AI caught a problem at which layer?
Correct. The AI identified that a newsworthy claim was buried in the third section — a structural, developmental edit — which a human sub-editor had missed.
The AI's observation was structural: the placement of key information, not surface errors or grammar.
3. In "diagnostic mode," what is the writer asking AI to do?
Correct. Diagnostic mode separates problem identification from fixing — putting the model in a structural frame before any line-level rewriting begins.
That describes generative or comparative mode. Diagnostic mode focuses on naming problems, not solving them.
4. What is "anchoring bias" in the editing context?
Correct. Anchoring bias means the more times you read your own draft, the harder it becomes to perceive its flaws — one reason a fresh reader (or AI) can catch things you miss.
Anchoring bias is the human cognitive tendency to over-weight the first version encountered, which makes self-editing increasingly difficult over time.

Lab 1 — Diagnostic Editorial Session

Practice asking AI for a structural diagnosis before any fixes

Your Mission

Paste a short passage (3–8 sentences) of any draft writing — fiction, journalism, essay, anything. Then ask the AI editor to diagnose its structural problems before suggesting any rewrites. Practice separating diagnosis from prescription.

Try prompts like: "What are the two biggest structural problems with this passage?" or "Where does a reader lose the thread here?"

Complete 3 exchanges to finish the lab. Focus on getting the AI to tell you what is wrong before asking it to fix anything.
AI Story Editor
Diagnostic Mode
Welcome to the editorial lab. Paste a passage you're working on — even a rough draft — and ask me to diagnose its structural issues. I'll tell you what I see before we discuss any fixes. What are you working on?
Module 5 · Lesson 2

Pacing, Tension, and the Reader's Pulse

How AI can identify where a story accelerates, stalls, or loses momentum — and what to do with that information.
Can a machine sense that a story is boring? What does "pacing" actually look like in data terms?

In 2023, Joanna Maciejewska, a fantasy novelist, published detailed notes on her Substack about using Claude to map tension curves in her manuscript. She asked the model to rate each chapter's tension on a 1–10 scale and explain its reasoning. The resulting chart revealed that three consecutive middle chapters all scored 3–4 — a sag she had felt but could not precisely locate. She later called it "the most useful single piece of editing feedback I got on that book."

What Pacing Actually Is

Pacing is not synonymous with speed. A slow scene can have intense forward momentum; a fast-action sequence can feel oddly inert. Pacing is the reader's felt sense that something is at stake and changing — that the story is earning their attention moment to moment.

The craft elements that generate pacing include: sentence length variation, scene-to-summary ratio, the density of new information per paragraph, unresolved questions (micro-tension), and the distance between a dramatic question being posed and answered.

How AI Reads Pacing

AI does not feel boredom. But it can identify proxies for the conditions that produce boredom in readers. Specifically:

Information density drops: When paragraphs contain mostly re-statement of what was already known, rather than new facts, questions, or complications.

Sentence length homogeneity: Prose that runs at a consistent medium length for many consecutive sentences tends to lull — the rhythm stops surprising.

Scene without consequence: Extended scenes where no character decision or revelation changes the situation. The reader senses the story has paused to describe rather than to act.

Unresolved question drought: Strong pacing keeps open questions in the air simultaneously. When all questions from a prior scene are answered and no new ones are raised, momentum stalls.

Research Context

A 2023 analysis by researchers at the Allen Institute for AI (AI2) found that GPT-4 could identify "tension arcs" in short fiction at roughly 74% agreement with trained human editors when prompted with structured evaluation criteria. Without structured prompts, agreement dropped to 41% — a useful reminder that prompt quality determines output quality.

The Tension Mapping Protocol

  • 1Break your manuscript into units (chapters, scenes, or sections of ~300–500 words).
  • 2Give AI each unit and ask: "Rate the tension/forward momentum of this passage 1–10, and explain the two specific things driving that score."
  • 3Chart the scores. Look for plateaus (three or more consecutive low scores) and cliffs (sudden unexplained drops).
  • 4For plateau sections, ask: "What open question could be introduced here that would give the reader a reason to continue?"
  • 5For cliff sections, ask: "What did the reader just lose that created this drop — information, relationship, stakes?"
The Novelist's Caveat

Zadie Smith, in a 2023 interview with The Guardian, noted that some of her favourite passages in her own work would score poorly on any tension metric — they are deliberately ruminative. The point of tension mapping is not to eliminate slow passages but to ensure your slow passages are intentional choices, not invisible problems.

Micro-tensionSmall, unresolved questions or uncertainties that keep a reader engaged even in quiet scenes — distinct from the macro-tension of plot stakes.
Scene-to-summary ratioThe proportion of dramatised real-time action versus narrative summary in a passage. High summary often correlates with low reader engagement.
Tension plateauThree or more consecutive sections scoring similarly low on a tension mapping exercise — a structural signal that the story has stopped escalating.

Lesson 2 Quiz

Pacing, Tension, and the Reader's Pulse — 4 questions
1. What is a "tension plateau" in the context of AI pacing analysis?
Correct. A tension plateau is three or more consecutive sections scoring low — a structural signal the story has stopped escalating.
A plateau specifically means multiple consecutive low-scoring sections, not a single scene or the story's peak.
2. According to the AI2 research cited in this lesson, what dramatically improved AI-human agreement on tension identification?
Correct. With structured prompts, human-AI agreement rose to 74%. Without them it dropped to 41% — demonstrating that prompt quality is critical.
The key variable was structured prompts. Without them, agreement fell to 41%, regardless of model size or passage length.
3. Which of these is NOT a proxy AI uses to identify low pacing?
Correct. Emotional authenticity is precisely the kind of judgment AI cannot reliably make — it requires lived human experience to assess.
Emotional authenticity is the outlier here. AI can track information density, sentence variation, and open questions — but cannot genuinely evaluate whether a character's emotion feels true.
4. What was Joanna Maciejewska's key finding when she used AI to map her manuscript's tension?
Correct. The tension chart revealed a specific three-chapter plateau she had sensed but couldn't isolate — she called it the most useful editorial feedback on the book.
The AI provided diagnostic feedback — specifically identifying a three-chapter tension plateau — not rewrites or grammar corrections.

Lab 2 — Tension Mapping

Use AI to score and analyse the pacing of your own writing

Your Mission

Paste a scene or passage (at least a paragraph) and ask the AI to score its tension 1–10 and explain the specific factors driving that score. Then experiment with what the AI suggests would raise or lower the score.

Try: "Rate this passage's pacing 1–10 and name the two specific things most responsible for that score."

Complete 3 exchanges to finish the lab. Push past the first score — ask follow-up questions about what would change it.
AI Story Editor
Tension Mapping
Tension mapping lab is open. Share a passage and I'll score its pacing and forward momentum on a 1–10 scale, explaining exactly what's driving that score. Then we can explore how to adjust it. What do you have?
Module 5 · Lesson 3

Voice Preservation: Editing Without Erasure

The most common complaint about AI editing is that it smooths everything into the same beige sameness. Here is how to prevent it.
How do you use AI as an editor without letting it replace your voice with its own?

In October 2023, literary magazine Granta published an editorial by their digital team documenting an internal experiment: they ran 30 submissions through GPT-4 for a "clarity pass" before human editorial review. When the human editors were shown the AI-touched manuscripts without being told which had been processed, they rated the AI-edited prose as less distinctive on average — noting a flattening of idiosyncrasy, unusual syntax, and deliberate rhythm breaks. Granta's conclusion: AI copy-editing, applied without constraint, optimises for readability at the cost of voice.

What "Voice" Actually Consists Of

Voice is not personality sprinkled on top of prose. It is a set of specific, consistent micro-decisions: sentence rhythm, the ratio of concrete to abstract, when the writer interrupts themselves, how they handle transitions, what they never explain, what they over-explain, the particular distance they maintain from their subject.

AI models, particularly when instructed to "improve" or "clarify," default toward consensus patterns in their training data — which is to say, toward what most edited prose looks like. That means longer sentences get broken, unusual sentence openings get normalised, and the choices that make a writer recognisable get quietly removed.

The Voice-Lock Technique

The solution is to explicitly teach the model your voice before asking it to edit. This has three practical steps:

Step 1 — Voice sample: Provide 3–5 paragraphs you consider to be strong examples of your best writing. Tell the AI: "This is my voice at its clearest. Study the sentence rhythm, the way I handle transitions, the level of formality."

Step 2 — Voice inventory: Ask the AI to list the specific stylistic features it observes in your sample. Review its list. Add anything it missed; correct anything wrong.

Step 3 — Constrained edit: Only now submit the passage needing editing, with the instruction: "Edit for [specific issue] without changing any of the voice features we identified."

Case in Point

In 2023, essayist and critic Hanif Abdurraqib described his use of AI editing tools in an interview with Vulture: "I use it like I'd use a cold read from someone who doesn't know me yet. I give it a lot of context about what I'm trying to do. Without that context, it tries to make everything sound like everything else."

What to Protect vs. What to Surrender

Protect These

Sentence fragments used deliberately. Unusual punctuation rhythm. Unexpected tonal shifts. Repetition used as a device. Sentences that break grammar rules for effect. Any construction that appears repeatedly — it is likely a signature.

Surrender These

Accidental repetition (a word used three times in one paragraph without intent). Unclear antecedents. Tense inconsistencies. Paragraphs that could be read two ways but you only intend one. Real typos, not stylistic ones.

The Editor's Principle

The best human editors have always worked by this principle: the job is to help the writer be more themselves, not to impose an external standard. That principle applies equally to AI editorial work. The difference is that a human editor unconsciously absorbs your voice after reading your work. AI requires you to make that context explicit.

Voice-lockThe technique of providing an AI editor with explicit voice samples and a feature inventory before requesting edits, to prevent stylistic flattening.
Readability optimisationAI's default tendency to make prose clearer and more conventionally fluent — useful for access, but potentially destructive to distinctive style.
Stylistic signatureA recurring micro-decision in a writer's prose that, taken together with others, constitutes their recognisable voice.

Lesson 3 Quiz

Voice Preservation — 4 questions
1. What did Granta's 2023 internal experiment find about AI copy-editing?
Correct. Human editors rated AI-edited prose as less distinctive on average, noting a flattening of idiosyncrasy and unusual syntax.
The finding was the opposite: the AI-touched manuscripts were rated as less distinctive, not better or undetectable.
2. What is the correct order of the voice-lock technique?
Correct. You establish voice first (sample), confirm the model's understanding (inventory), then request the edit with those constraints active.
Voice-lock requires establishing voice before editing begins — not after, and not by asking the AI to infer it from the passage you want edited.
3. According to this lesson, which of these should a writer typically PROTECT from AI editing?
Correct. Deliberate fragments are a stylistic signature — protecting them is part of preserving voice. The other options are genuine errors, not style choices.
Tense inconsistencies, unclear antecedents, and accidental repetition are genuine problems to surrender. Deliberate fragments are voice — protect them.
4. Why does AI default toward "readability optimisation" rather than preserving distinctive voice?
Correct. Training on consensus-edited prose means the model gravitates toward what most edited writing looks like — which erases the outlier choices that constitute individual voice.
AI can process unusual syntax. The issue is that its training data skews toward conventionally edited prose, so it gravitates toward those patterns when "improving" writing.

Lab 3 — Voice-Lock Workshop

Practice establishing your voice before asking for edits

Your Mission

Use the voice-lock technique: first share a strong writing sample and ask the AI to identify your stylistic signatures. Then submit a different passage for editing with those signatures explicitly protected.

Start with: "Here is a writing sample I'm proud of. Please identify the specific stylistic features that define my voice."

Complete 3 exchanges. The goal is to get the AI to name your voice features, then use them as a constraint on an edit request.
AI Story Editor
Voice Preservation
Voice-lock lab. Start by sharing a writing sample you consider representative of your best work. I'll identify the specific stylistic signatures I observe — sentence rhythm, vocabulary range, structural habits — and we'll use that inventory to protect your voice in any subsequent editing. Ready when you are.
Module 5 · Lesson 4

The Revision Loop: Iterating with AI Without Losing the Thread

How to manage multiple revision passes, track what has changed, and know when to stop — the discipline of AI-assisted revision.
When does AI-assisted revision become revision paralysis — and how do you build a process that ends?

In February 2024, The Atlantic published a reported essay by Ian Bogost examining the phenomenon of writers becoming trapped in endless AI revision cycles. Bogost interviewed a cohort of freelance journalists and found that several had used AI editing tools to revise pieces 15–20 times before filing — compared to 3–5 passes previously. The pieces were not measurably better; in two documented cases, editors at the publications noted that the final filed drafts felt over-worked — the prose had lost spontaneity. The problem was not AI but the removal of any forcing function to stop.

The Revision Trap

AI-assisted revision introduces a specific new failure mode: because revision is now nearly frictionless, writers can revise indefinitely. Every pass surfaces new suggestions. Every suggestion, implemented, creates new micro-inconsistencies that the next pass finds. The draft is always improvable. The loop never ends.

Traditional editorial workflow had natural stopping points built in: a deadline from an editor, the physical cost of retyping, the scarcity of trusted readers. AI removes all three. Without a replacement structure, writers drift.

Building a Revision Protocol That Ends

  • 1Define passes before you begin. Decide: Pass 1 is structural diagnosis only. Pass 2 is pacing. Pass 3 is voice and line-level. Pass 4 is final copy-edit. No pass may address a prior pass's concerns.
  • 2Write a "done" criteria statement. Before you open the draft, write: "This piece is finished when [specific criteria]." Keep it visible. AI suggestions that do not address those criteria are declined.
  • 3Maintain a change log. After each AI-assisted pass, write one sentence summarising what changed and why. If you cannot summarise it, reverse the change.
  • 4Version-lock at submission. Rename the file with a date stamp and stop. Any further suggestions from AI sessions go into a separate "future draft" document.
Workflow Evidence

In 2023, the content team at Mailchimp published internal guidelines on AI-assisted editing that included a mandatory "revision ceiling" — no more than four AI-assisted passes per piece. Team leads reported a 30% reduction in the time-to-publish for long-form content without editors reporting quality degradation. The ceiling was the key variable.

Recognising When a Draft Is Done

AI cannot tell you a piece is finished. It can always find something to refine. The judgment that a piece has achieved its purpose is a human editorial call, and a crucial one.

Signals that a draft is done, despite AI suggestions to the contrary: the piece does what you set out to do; every remaining AI suggestion addresses a feature you want rather than need to change; you have implemented all changes you can justify; you would not be embarrassed to have the work published in its current state.

The novelist Anne Lamott wrote in Bird by Bird (1994) about the concept of "good enough" — the moment when further revision improves the piece technically but no longer brings it closer to its essential intention. That principle predates AI but applies with new urgency to AI-assisted revision workflows.

The Core Principle

AI is an instrument for serving your editorial vision. When you find yourself implementing AI suggestions without being able to articulate why they improve the piece toward your stated goals, the tool has inverted the relationship — and it is time to stop.

Revision ceilingA pre-determined maximum number of AI-assisted revision passes, set before work begins, to prevent indefinite refinement.
Done criteriaA written statement of what the piece must accomplish to be considered finished, used to evaluate whether AI suggestions are worth implementing.
Over-workingThe condition in which excessive revision has removed spontaneity and naturalness from prose — technically improved but experientially diminished.

Lesson 4 Quiz

The Revision Loop — 4 questions
1. What specific failure mode does AI-assisted revision introduce that traditional editing did not?
Correct. Traditional workflows had built-in friction — deadlines, physical retyping costs, reader scarcity. AI removes all of these, enabling indefinite revision loops.
The failure mode is structural, not cognitive: AI eliminates the friction that used to force stopping points in the revision process.
2. What did Ian Bogost's Atlantic essay find about AI-assisted revision practices among freelance journalists?
Correct. Bogost found writers doing 15–20 passes with AI editing tools, and editors in some cases noted the results felt over-worked — more revision did not mean better work.
The finding was that more revision passes with AI did not produce better pieces — and in documented cases, editors found the prose felt over-worked.
3. What was the key variable in Mailchimp's 2023 AI editing workflow that reduced time-to-publish by 30%?
Correct. The revision ceiling — a hard limit of four AI-assisted passes — was identified as the key variable that reduced time-to-publish without quality degradation.
The specific intervention was a revision ceiling: a pre-set maximum number of passes that forced decisions and prevented infinite refinement loops.
4. According to this lesson, when should a writer STOP implementing AI suggestions?
Correct. The signal to stop is when AI suggestions have no clear relationship to your stated purpose — at that point, the tool has inverted the relationship between writer and instrument.
AI will always find something to refine. The stopping signal is internal: when you can no longer articulate why a suggested change brings the piece closer to its stated purpose.

Lab 4 — Structured Revision Protocol

Practice designing and executing a capped, purposeful revision pass

Your Mission

Write a "done criteria" statement for a piece you're working on, then run one focused revision pass with a specific goal. Practice declining AI suggestions that don't serve your stated criteria.

Start with: "Here are my done criteria for this piece: [your criteria]. Now review this passage for [one specific issue only] and ignore everything else."

Complete 3 exchanges. Practise the discipline of saying "that suggestion doesn't serve my criteria — skip it." The goal is a revision pass that has a defined end.
AI Story Editor
Revision Protocol
Structured revision lab. Before we edit anything, tell me your "done criteria" — what does this piece need to accomplish to be finished? Once we've agreed on that, submit a passage and tell me which single issue you want this pass to address. I'll stay in that lane and not offer suggestions outside your brief.

Module 5 — Module Test

AI as Story Editor · 15 questions · Pass at 80%
1. What is the highest-order layer of editing addressed in this module?
Correct. Developmental editing — concerned with structure, argument, and pacing — is the highest-order layer, and the one where AI offers the most surprising editorial value.
Proofreading and copy-editing work at the surface and sentence levels. Developmental editing is the deepest layer.
2. In "comparative mode," what does the writer ask AI to do?
Correct. Comparative mode uses two versions and asks which better serves a specific stated purpose — useful when you're caught between revisions.
Comparative mode specifically involves two versions evaluated against a goal. Diagnosing problems is diagnostic mode; demonstrating alternatives is generative mode.
3. What unique advantage does AI have over humans when performing developmental editing on a long document?
Correct. AI can compare paragraph 12 against paragraph 2 without fatigue or the anchoring bias humans develop after multiple re-readings.
AI's structural advantage is holding the whole document without fatigue or anchoring bias — not factual knowledge or emotional judgment.
4. What are "micro-tension" and "macro-tension" respectively?
Correct. Micro-tension refers to small open questions that keep readers engaged in otherwise quiet scenes; macro-tension is the larger stakes of the plot or argument.
Micro-tension is small open questions within a scene; macro-tension is the large-scale plot stakes — not a sentence-length or structural-level distinction.
5. The Janklow & Nesbit literary agency experiment in 2023 found that AI was particularly useful for identifying what?
Correct. Agents found the tool consistently identified character motivation inconsistencies — catches that previously required a second human reader.
The agency's finding was about structural consistency — specifically character motivation established inconsistently across chapters.
6. What does a tension mapping exercise chart specifically?
Correct. Tension mapping charts AI-rated momentum scores section by section, revealing plateaus (multiple low scores) and cliffs (sudden unexplained drops).
Tension mapping uses section-by-section AI scoring to reveal where momentum stalls or drops — not word counts, dialogue ratios, or sentence lengths directly.
7. According to the AI2 research, what happened to human-AI agreement on tension identification without structured prompts?
Correct. Without structured prompts, human-AI agreement on tension identification fell from 74% to 41% — a stark demonstration of how much prompt design matters.
The drop was significant: from 74% with structured prompts to 41% without — showing that prompt quality is a major variable in AI editorial work.
8. What is a "stylistic signature" as defined in Lesson 3?
Correct. Stylistic signatures are recurring micro-decisions — taken together, they constitute what makes a writer recognisable and what must be protected during AI editing.
A stylistic signature is a recurring micro-decision in the writer's own prose — not a label, program technique, or AI output.
9. What is the correct first step of the voice-lock technique?
Correct. The first step is to provide voice samples and let AI generate its own observations — you then review and correct that inventory before any editing begins.
Voice-lock starts with providing voice samples so AI can identify features — not editing first, not asking for a bio-based description, and not listing features yourself without AI confirmation.
10. Granta's 2023 experiment specifically found that AI copy-editing, applied without constraint, optimised for what at the expense of what?
Correct. Granta found that unconstrained AI copy-editing improved surface readability while flattening the idiosyncrasy and unusual syntax that make writing distinctive.
The trade-off Granta identified was readability versus voice — not speed, grammar, or concision.
11. What is "over-working" in the context of AI-assisted revision?
Correct. Over-working is the editorial condition where too many revision passes have polished away the natural energy of the prose, even as grammar and structure improve.
Over-working is a quality problem — prose that has been revised so many times it feels lifeless, despite being technically correct.
12. Ian Bogost's Atlantic essay documented freelance journalists completing how many AI-assisted revision passes before filing?
Correct. Bogost found some journalists doing 15–20 revision passes with AI tools, compared to 3–5 previously — with no corresponding quality improvement.
Bogost specifically documented 15–20 passes for some journalists, a dramatic increase from the pre-AI 3–5 passes, without measurable quality improvement.
13. What is a "revision ceiling" and why is it important?
Correct. A revision ceiling is a structural constraint the writer sets before beginning — a hard limit on passes that creates the forcing function AI-assisted workflows otherwise lack.
A revision ceiling is a self-imposed structural limit on revision passes — not a publisher constraint, AI capability limit, or software setting.
14. Which of the following is NOT a signal that a draft is done despite remaining AI suggestions?
Correct. AI will always generate new suggestions — that is not a signal that you are done. The signals that matter are internal: purpose achieved, changes unjustifiable, publication-ready.
AI doesn't stop generating suggestions — "the AI ran out of ideas" is not a stopping signal. The signals are all about your editorial judgment, not the tool's output.
15. What is the core principle governing the relationship between writer and AI editor, as stated in Lesson 4?
Correct. The core principle is that AI serves your vision. When suggestions no longer serve your stated goals and you cannot articulate why to implement them, the tool has inverted the relationship.
The core principle is about maintaining the writer's editorial authority — AI as instrument, not authority. More passes do not mean better work; unjustifiable suggestions should be declined.