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

What "Voice" Actually Means

The difference between words on a page and a person on the page.
How do readers know it's you — and not just text?

In 2023, literary agents reported receiving floods of AI-assisted query letters — hundreds of proposals for books on identical trending topics, using nearly identical phrasing. Most were rejected within seconds. One agent, Janet Reid of New Leaf Literary, noted publicly that the difference between the slush pile and a request for pages was almost always a single factor: whose specific perspective made the idea irreplaceable.

Voice isn't vocabulary. It's the accumulation of choices — what you notice, what you skip, how fast you move, when you're funny, when you're still — that makes a reader feel they are with a particular human being.

The Three Layers of Voice

Most writers think of voice as style — word choice, sentence length, whether you use contractions. That's the surface layer and it's the easiest for AI to approximate. Below it are two deeper layers that AI cannot replicate because they require lived experience.

Surface Layer (AI can mimic)

Diction, sentence rhythm, punctuation habits, formality level, use of humor, paragraph length, transition words. These patterns can be learned from examples and reproduced statistically.

Deep Layers (AI cannot source)

The specific things you've noticed, the opinions you actually hold, the comparisons only you would make, the gaps you leave because you know the reader already knows — these come from living a particular life.

What Makes Voice Recognizable

Voice researchers and writing teachers have studied this for decades. The short answer: voice is recognized by specificity and omission. What the writer chooses to include — and what they trust themselves to leave out — reveals perspective more than any stylistic flourish.

Consider two descriptions of the same event:

Generic (no voice)

"The concert was amazing. The crowd was excited. The music was loud and the performance was energetic."

With Voice

"The bass hit somewhere in my sternum. The woman next to me had been crying since the opening chord and she didn't care who saw. I've been to maybe forty concerts. This was the one I'll describe to my grandchildren."

The second version assumes things: you know what a sternum is, you've had the "this is the one" feeling before. It omits things: it never says the genre, the venue, the artist's name. That trust — in both writer and reader — is voice.

Why AI Defaults to Generic

AI language models are trained on the statistical center of human writing. Their outputs trend toward the most commonly co-occurring word sequences — which means toward averaged, conventional expression. When you ask an AI to "write in my voice" without guidance, it produces the voice of Everyone, which is the voice of No One.

This is not a flaw to be fixed. It is the architectural reality of how these models work. Understanding it lets you use AI precisely where it helps (generating raw material, varying structure, filling in facts) while keeping yourself in the seat where only you belong: deciding what the piece is actually saying.

VoiceThe cumulative effect of a writer's choices — what to include, omit, emphasize, and trust — that makes a reader feel the presence of a specific person.
Statistical CenterThe AI tendency to produce the most average, commonly-expected word sequences, resulting in writing that sounds competent but generic.
SpecificityThe use of precise, personal detail that no one else would have chosen — the primary marker of authentic voice.
The Core Principle

AI can write words. Only you can write your words. The distinction is not about effort — it's about origin. Your specific perspective, formed by your specific life, is the only resource in this workflow that is genuinely non-renewable.

Quiz — Lesson 1

What "Voice" Actually Means
1. According to literary agents, what primarily separated strong query letters from the AI-assisted slush pile in 2023?
Correct. Janet Reid and other agents cited unique personal perspective — not polish — as the deciding factor.
Not quite. Agents flagged that AI-assisted letters often had fine grammar but lacked a specific human perspective worth reading.
2. Which layer of voice can AI most readily approximate?
Correct. Surface stylistic patterns are statistical and learnable. Deeper layers require lived experience.
Surface style — diction, rhythm, punctuation habits — is the layer AI can mimic from examples. The deeper layers require lived experience.
3. Voice is primarily recognized by which two qualities?
Correct. What a writer chooses to include (specificity) and what they trust themselves to leave out (omission) are the primary markers.
The lesson identifies specificity and omission — what you include and what you trust yourself to skip — as the primary markers of voice.
4. Why does AI tend toward generic expression rather than distinctive voice?
Correct. Training on vast text averages toward the most common co-occurring word sequences — the voice of Everyone, which is the voice of No One.
The architectural reason is statistical: models produce the most commonly co-occurring word sequences, which averages toward the conventional center.
5. In the concert example, what technique most strongly creates voice?
Correct. "Bass hit in my sternum," the unnamed crying woman, forty concerts of context — plus trusting the reader without naming the artist — creates a felt presence.
The concert example works through specific sensory/emotional detail (bass in the sternum, the crying woman) and strategic omission (no artist name) — not technical description or completeness.

Lab 1 — Voice Anatomy

Identify and articulate the specific elements of your own voice

Your Voice Inventory

Before you can protect your voice in AI-assisted writing, you need to know what it actually consists of. In this lab, you'll work with an AI coach to build a personal voice inventory — the specific markers that make your writing yours.

The AI will ask you questions about your writing habits, preferences, and tendencies. Answer honestly and specifically. After 3+ exchanges, you'll have a working voice profile to use in later labs.

Start by telling the AI: what topic do you write about most often, and what's one thing you always notice that other writers seem to miss?
Voice Coach
Lab 1
Welcome to your Voice Inventory. I'm going to help you map the specific elements that make your writing yours — not in the abstract, but with concrete markers you can actually use. Start by telling me: what topic do you write about most often, and what's one thing you always notice that other writers seem to miss?
Module 5 · Lesson 2

How AI Flattens Voice

Understanding the specific mechanisms that strip distinctiveness from your writing.
What exactly happens to your writing when you hand it to AI — and why?

In February 2023, Clarkesworld Magazine — one of science fiction's most prestigious short fiction venues — temporarily closed submissions after the volume of AI-generated stories overwhelmed its slush pile. Editor Neil Clarke reported that the number of spam submissions (his term for AI-generated stories without disclosed AI use) had risen from single digits monthly to hundreds within weeks.

What made the AI stories identifiable wasn't that they were bad in the conventional sense. They were competent. They were readable. They had beginnings, middles, and ends. What they lacked was any sense that a particular person needed to write this particular story. They felt, Clarke noted, like stories produced by someone who had read about stories rather than lived through experiences that demanded expression.

The Five Flattening Mechanisms

AI doesn't just remove bad writing. It removes the eccentric writing — the unexpected choices, the jarring turns, the moments where a writer's specific sensibility overrides conventional technique. Here are five specific mechanisms through which this happens:

1.
Regression to the Mean Phrase. When your draft uses an unusual word or construction, AI editing tends to replace it with the statistically more common alternative. "The meeting adjourned into mutual exhaustion" becomes "The meeting ended and everyone was tired."
2.
Smoothing of Cadence. Unusual sentence rhythms — very short sentences after long ones, fragments used deliberately, run-ons that mirror thought — get normalized into more even, conventional prose flow.
3.
Deletion of Personal Asides. The parenthetical remark, the digression, the "I should mention" — these are where personality lives. AI often removes them as "off-topic" or restructures them into the main argument, eliminating the texture.
4.
Generic Conclusion Insertion. AI tends to add summary sentences that restate what was just said. These feel reassuring to a model but to a reader, they signal that the writer didn't trust them — which kills intimacy.
5.
Hedging and Qualification. Strong voice often makes bold, unhedged claims — "This is wrong," not "Some might argue this could potentially be seen as problematic in certain contexts." AI tends to add qualifications that soften claims into inoffensiveness.
The Competence Trap

The most insidious effect of AI on voice is not that it makes writing worse. It makes writing professionally adequate — which, for most creative and persuasive purposes, is worse than bad. Bad writing at least signals a human trying. Adequate writing is forgettable.

In 2023, researchers at the University of Pennsylvania's Wharton School found that when business writing was AI-edited, evaluators consistently rated it as clearer but less memorable and less persuasive. The editing improved surface clarity while removing the distinctive assertions that made readers remember and act on the content.

Real Data Point

In the Wharton study, AI-edited memos scored 23% higher on "clarity" ratings but 31% lower on "I would act on this" ratings compared to original human drafts. Clarity and persuasiveness moved in opposite directions.

What You're Actually Losing

When AI flattens your voice, the specific losses are: the reader's sense of your credibility (because your unusual observations prove you've actually thought about the subject), the emotional investment that comes from feeling a real person wrote this, and the memorability that comes from encountering something you hadn't seen that way before.

None of these can be added back by further AI editing. They have to be preserved during the editing process — which means knowing, precisely, what you're protecting before you start.

Regression to Mean PhraseAI's tendency to replace unusual word choices with statistically more common alternatives, eliminating the specificity that creates voice.
Competence TrapThe condition in which AI editing produces writing that is technically adequate but forgettable — professionally safe but persuasively inert.
Cadence SmoothingThe normalization of unusual sentence rhythms — fragments, run-ons, extreme length variation — into conventional, even prose flow.

Quiz — Lesson 2

How AI Flattens Voice
1. What was the primary reason Clarkesworld identified AI-generated submissions as problematic?
Correct. Editor Neil Clarke described the AI stories as readable and complete — but without any felt necessity behind them.
The issue wasn't quality in the conventional sense — the stories were technically fine. They lacked the felt necessity of a specific person who needed to tell that story.
2. Which mechanism describes AI replacing "the meeting adjourned into mutual exhaustion" with "the meeting ended and everyone was tired"?
Correct. Replacing an unusual, specific construction with a statistically more common equivalent is regression to the mean phrase.
This is regression to the mean phrase — replacing an unusual, distinctive construction with a more statistically common alternative.
3. What did the Wharton School research find about AI-edited business writing?
Correct. AI-edited memos scored higher on clarity but 31% lower on "I would act on this" — clarity and persuasiveness moved in opposite directions.
The Wharton finding was that AI editing improved clarity ratings while reducing persuasiveness — the two qualities moved in opposite directions.
4. Why is "professionally adequate" writing described as worse than bad for creative and persuasive purposes?
Correct. The competence trap is that forgettable adequate writing achieves less than imperfect writing with a genuine human sensibility behind it.
The competence trap: adequate writing is forgettable. Bad writing at least signals a human trying. For persuasion and creative purposes, forgettable is the worst outcome.
5. Personal asides and digressions are described as important because:
Correct. The parenthetical, the digression, the "I should mention" — these are the sites of felt personality, not distractions from it.
Personality lives in the aside and the digression — not in the main argument. When AI removes them as "off-topic," it removes the texture that makes readers feel a person wrote this.

Lab 2 — Flattening Detection

Spot the specific places where AI removed your voice

Before and After Analysis

In this lab, you'll paste a short piece of your own writing (3–6 sentences) and an AI-edited version of it. The AI coach will identify exactly which of the five flattening mechanisms occurred and what was lost.

If you don't have an AI-edited version, paste just your original writing and ask the AI to "edit it to improve clarity" — then analyze what changed.

Paste your original text and the AI-edited version (or just your original text and say "please edit for clarity first"). Then ask: "What specific voice elements were removed?"
Voice Analyzer
Lab 2
Welcome to the Flattening Detection lab. Paste your original writing — 3 to 6 sentences — and either an AI-edited version alongside it, or just the original and I'll edit it first so we can compare. I'll identify exactly which flattening mechanisms occurred and what specific voice elements were lost or changed.
Module 5 · Lesson 3

Protecting Voice During AI Editing

Specific techniques to use AI's power without losing your distinctiveness.
How do you let AI improve your work without letting it replace your perspective?

In September 2023, George R.R. Martin joined a class-action lawsuit against OpenAI, alleging his distinctive voice and narrative style had been used without permission to train language models. The legal argument aside, the case illuminated something technically significant: Martin's voice was recognizable enough to be identified, documented, and claimed as property.

The characteristics cited were specific: his use of dramatic chapter-ending revelations, his management of point-of-view across ensemble casts, his particular mixture of high-register archaism and colloquial brutality. These weren't vague stylistic impressions — they were enumerable, teachable elements. If you can enumerate your voice, you can protect it.

The Protection Framework: MARK, LOCK, RESTORE

Voice protection in AI-assisted editing requires a three-stage process. Each stage addresses a different vulnerability window.

📌
MARK — Before You Submit to AI. Before pasting your draft into any AI tool, explicitly identify your "non-negotiables" — the specific phrases, rhythms, or structural choices that are intentional and must not be changed. Write them in the prompt: "Do not change: [list them]."
🔒
LOCK — In the Prompt Instructions. Specify what kind of help you want, not open-ended "improve this." Examples: "Fix only grammar errors," "Suggest alternative phrasings for the third sentence only," "Identify factual errors but do not rewrite prose." Task-specific prompts dramatically reduce unintended voice replacement.
🔄
RESTORE — After Receiving the Output. Compare the AI output to your original systematically. For each change, ask: "Did AI make this better or did it make it blander?" Restore any phrase that was distinctive and hasn't been made clearer, only different.
Prompt Engineering for Voice Preservation

The most effective prompt structure for voice-preserving AI editing has three components:

Voice-Preserving Prompt Structure

1. Role definition: "You are a copy editor, not a ghostwriter."
2. Constraint list: "Do not change sentence rhythms, do not soften claims, do not remove personal observations."
3. Specific task: "Fix only: subject-verb agreement errors, inconsistent tense, and unclear pronoun references."

This is dramatically more effective than "edit this for me" — which gives AI the latitude to apply all five flattening mechanisms. The more specific your constraint list, the more your voice survives the edit.

Using Comparison Drafts

A highly effective technique: ask AI to produce three alternatives to a specific sentence rather than a complete rewrite of a paragraph. Then compare all three to your original and cherry-pick — or use the alternatives to spark your own revision. You remain the decision-maker; AI generates options.

This approach was documented in a 2023 study by Stanford's Human-Computer Interaction group, which found that writers who used AI to generate options rather than replacements reported significantly higher satisfaction with final drafts and rated the work as more authentically theirs.

The "Explain, Then Edit" Protocol

Before asking AI to make any change, ask it to explain what it would change and why. This creates a review layer where you can accept the reasoning or reject it before the text is altered. It also trains you to recognize which of your habits are genuine voice choices versus genuine errors — a distinction that's harder to see in your own work.

Practical Rule

If an AI edit makes your writing sound like it could have been written by anyone with your rough topic knowledge, it has removed your voice. If it makes your writing sound more precisely like the thing you were trying to say, it has helped. The test is always: whose thought is this now?

MARK / LOCK / RESTOREA three-stage voice protection framework: identify non-negotiables before submitting, constrain the AI task in the prompt, and systematically restore distinctive elements afterward.
Options ModeUsing AI to generate multiple alternatives to a single element rather than a full rewrite, preserving the writer as decision-maker.
Explain Before EditAsking AI to describe what it would change and why before executing the change, creating a review layer that protects intentional choices.

Quiz — Lesson 3

Protecting Voice During AI Editing
1. What did the George R.R. Martin lawsuit illustrate about voice protection?
Correct. The case's significance was that Martin's voice characteristics were specific enough to enumerate — which means specific enough to protect intentionally.
The legal outcome aside, the case showed that when voice elements are specific and enumerable — Martin's POV management, his archaism mixed with colloquialism — they can be identified and documented. That's the protection insight.
2. In the MARK / LOCK / RESTORE framework, what does the LOCK stage address?
Correct. LOCK happens in the prompt — specifying exactly what task you want performed so AI doesn't have latitude to apply all five flattening mechanisms.
MARK is pre-submission identification of non-negotiables. LOCK is in the prompt — specifying the narrow task. RESTORE is the post-output comparison and repair stage.
3. According to Stanford HCI research, what editing approach left writers feeling the work was most authentically theirs?
Correct. Options mode — generating alternatives rather than replacements — preserved writer agency and produced higher authenticity ratings.
The Stanford finding was that asking AI for options (several alternatives to a specific sentence) rather than replacements (a full rewrite) produced significantly higher satisfaction and authenticity ratings.
4. What is the primary purpose of the "Explain, Then Edit" protocol?
Correct. By seeing the reasoning first, you can distinguish intentional voice choices from actual errors before any change is made.
The "Explain, Then Edit" protocol creates a review layer — you see the reasoning and can accept or reject it before the text is touched. This is especially useful for distinguishing intentional stylistic choices from actual mistakes.
5. The practical test for whether an AI edit has preserved or removed your voice is:
Correct. "Whose thought is this now?" — if it could have been written by anyone with your topic knowledge, voice has been lost.
The test is ownership of thought: if the edit makes the writing more precisely yours, it helped. If it makes the writing sound like anyone who knows the topic could have written it, voice has been removed.

Lab 3 — Constrained Editing Practice

Apply MARK / LOCK / RESTORE to a real piece of your writing

Voice-Protected Edit Session

In this lab, you'll practice the full MARK / LOCK / RESTORE workflow with a real piece of writing. Bring 1–3 paragraphs you've written. The AI will help you build a voice-preserving prompt, execute a constrained edit, and then identify what to restore.

The goal is not to get a perfect draft — it's to practice making the decisions that keep the work yours.

Start by pasting your paragraphs and listing 2–3 things in them that are intentional and should not change. Then say: "Help me write a LOCK prompt for editing only [specific problem]."
Edit Coach
Lab 3
Ready for the MARK / LOCK / RESTORE workflow. Paste your paragraphs and tell me 2–3 things in them that are intentional — choices you made that should survive any edit. Then tell me what specific problem you actually want fixed, and we'll build a constrained prompt that protects everything else.
Module 5 · Lesson 4

Adding Your Voice Back In

When AI writes the draft, your job is to inhabit it — completely and visibly.
How do you take something AI generated and make it unmistakably yours?

In late 2023, The New York Times released internal guidelines for AI-assisted journalism. One of the central requirements was that any AI-generated text must be substantially rewritten by the journalist before publication — not merely fact-checked, but transformed through the reporter's specific knowledge, sources, and perspective.

The Times editorial leadership framed this not as a restriction but as a definition: the journalist's job is not to verify AI output — it is to replace AI's generic framing with their own reporting and judgment. The AI draft is raw material, not a finished product requiring polish. The journalist's voice isn't added on top; it is the product itself.

The Four Injection Points

When working from an AI-generated draft, your voice has to be actively injected at four specific points. These are the sites where AI output is most generic and where your specificity has the highest impact:

🎯
The Opening Sentence. AI almost always opens with a statement of general context. Replace it with the most specific, surprising, or personal thing you know about the topic. This is the single highest-leverage location for voice in any piece.
💬
The Opinion Moments. Scan the AI draft for every place it hedges — "some argue," "it could be said," "many believe." Each one is an opportunity to replace the hedge with your actual opinion. If you don't have one, that's the signal to do more thinking before you publish.
🔍
The Examples and Evidence. AI examples are the most generic things in any draft — they're the first search-result illustrations. Replace them with examples from your own experience, reading, or reporting. Specific examples from lived knowledge are the strongest voice markers in expository writing.
The Closing Line. AI closings summarize and conclude. Replace the closing with the thing you most want the reader to carry away — and say it in the most specific language you can. Summaries are forgettable; specific final images and claims are not.
The Ratio Test

A practical benchmark used by professional writers working with AI: if fewer than 40% of the sentences in a final draft originated with you — either written by you originally or substantially transformed by you from AI output — the piece probably doesn't have your voice in it. It has your name on it.

This isn't a moral rule, it's a quality rule. The 40% threshold is roughly the minimum needed for your specific knowledge, opinions, and examples to dominate the reader's experience. Below it, the AI's generic framing tends to win.

The Name vs. Voice Distinction

Having your name on a piece and having your voice in it are different conditions. Your name is a byline. Your voice is the evidence that you — specifically — thought about this subject and found something particular to say. Readers can tell the difference, even when they can't articulate why.

Voice Injection Exercises

Professional writing coaches who work with AI-assisted drafts use a specific exercise: take an AI-generated paragraph and rewrite every sentence in it until you'd be comfortable defending each one in a conversation. "Defending" doesn't mean proving — it means that each sentence reflects your actual understanding, not just a plausible-sounding claim.

This exercise, practiced regularly, builds what voice coaches call sentence-level ownership — the habit of not publishing a sentence you couldn't explain in your own words from your own knowledge base.

The Culminating Principle

AI can build a scaffolding. You have to live in it. The scaffolding tells you where the rooms are; your voice tells the reader what it's like to be there. No amount of AI generation can substitute for a writer who has genuinely inhabited their subject — and the reader always knows the difference between a building with someone in it and an empty structure with lights left on.

Injection PointsThe four high-leverage locations in any AI draft where replacing generic content with your specific knowledge and opinions has the greatest impact on voice.
The 40% RuleA practical benchmark: fewer than 40% original or substantially transformed sentences typically means AI's generic framing dominates the reader's experience.
Sentence-Level OwnershipThe habit of not publishing a sentence you couldn't explain from your own knowledge and understanding — the baseline of voice authenticity.

Quiz — Lesson 4

Adding Your Voice Back In
1. How did The New York Times define the journalist's role when using AI-generated text?
Correct. The Times framed the journalist's role as transformation, not verification or polish — replacing generic AI framing with specific reporter judgment.
The Times guidelines defined the journalist's role as replacing AI's generic framing with their own reporting and judgment — not fact-checking or polishing AI output.
2. Why is the opening sentence described as the "single highest-leverage location for voice"?
Correct. AI default openings are statements of general context — maximally generic. A specific, surprising, or personal opener immediately differentiates the piece.
The leverage is because AI always defaults to general context openings — making any specific, surprising, or personal opener immediately differentiating. It's the highest-impact swap available.
3. When an AI draft uses "some argue" or "many believe," this signals an opportunity to:
Correct. Hedges are the AI's way of avoiding commitment. Each one is an injection point where your actual opinion should go — or a signal that you need to form one.
Hedging phrases are injection opportunities — places where AI avoided commitment and your actual opinion should go. If you don't have one, that's a signal to think more before publishing.
4. What does the 40% rule measure?
Correct. If fewer than 40% of sentences originated with or were substantially transformed by you, AI's generic framing likely dominates the reader's experience.
The 40% rule is a quality benchmark: if fewer than 40% of sentences are originally yours or substantially transformed by you, AI's generic framing tends to win over your voice.
5. "Sentence-level ownership" means:
Correct. Sentence-level ownership is the habit of not publishing anything you couldn't defend — explain from your own knowledge — in a conversation.
Sentence-level ownership is a habit of authorship: not publishing a sentence you couldn't explain from your own knowledge. It's about understanding and authority, not just writing the words yourself.

Lab 4 — Voice Injection Workshop

Transform an AI draft into something unmistakably yours

The Four Injection Points in Practice

In this lab, you'll work through the four injection points with an actual piece of writing. Either bring an AI-generated draft you want to transform, or ask the AI to generate a generic draft on a topic you know well — then work through replacing the opening, the hedges, the examples, and the closing with your own voice.

The AI coach will help you identify each injection point and evaluate whether your replacement is genuinely yours or still generic.

Paste an AI-generated draft (or a topic and ask for a generic draft to work from). Then say: "Identify the four injection points and evaluate my replacements as I make them."
Voice Injection Coach
Lab 4
Welcome to the Voice Injection Workshop. Paste an AI-generated draft you want to transform — or give me a topic you know well and I'll generate a deliberately generic draft for us to work from. Then we'll go through the four injection points: opening sentence, opinion moments, examples and evidence, and the closing line. I'll evaluate each replacement you make and tell you honestly whether it sounds like you or still like anyone.

Module 5 Test

Your Voice vs. AI's Voice — 15 questions · Pass at 80%
1. Voice is primarily the product of:
Correct. Voice is the accumulation of choices — what you notice, skip, trust — not any single stylistic feature.
Voice is the cumulative effect of choices about what to include and omit — not vocabulary range or stylistic signatures alone.
2. AI's "statistical center" tendency means it produces:
Correct. Training on vast text averages toward the statistically central — conventional, unremarkable expression.
The statistical center is the most commonly co-occurring sequences — not the best writing, but the most average writing.
3. Clarkesworld's 2023 crisis illustrated that AI-generated stories were problematic primarily because:
Correct. Editor Neil Clarke described the AI stories as readable and complete — but without felt necessity behind them.
The stories were technically competent — they lacked the felt necessity of a specific person who needed to tell that story.
4. Which of the following is an example of cadence smoothing?
Correct. Cadence smoothing targets unusual rhythms — fragments, run-ons, extreme length variation — and normalizes them into even prose.
Cadence smoothing normalizes unusual sentence rhythms. Replacing unusual word choices is regression to mean phrase; removing asides is deletion of personal asides; adding summaries is generic conclusion insertion.
5. The Wharton School study found that AI-edited business writing scored higher on clarity but lower on:
Correct. AI-edited memos scored 23% higher on clarity but 31% lower on "I would act on this" — persuasiveness and clarity moved in opposite directions.
The Wharton finding: clarity up, persuasiveness down. AI editing improved surface readability while removing the distinctive assertions that made readers act.
6. In the MARK / LOCK / RESTORE framework, MARK refers to:
Correct. MARK happens before you submit — explicitly listing what must not change, often directly in the prompt.
MARK is pre-submission: identifying and noting your non-negotiable choices before the draft ever reaches AI, often listed directly in the prompt.
7. A voice-preserving prompt should include which three components?
Correct. "You are a copy editor, not a ghostwriter" + constraints + specific task dramatically reduces unintended voice replacement.
The three components are: role definition ("copy editor, not ghostwriter"), constraint list (what not to change), and specific limited task (what to actually fix).
8. The "Explain, Then Edit" protocol creates value primarily by:
Correct. Seeing the reasoning lets you distinguish intentional voice choices from actual errors before any text is altered.
The protocol's value is the review layer: you see what AI would change and why, and can accept or reject the reasoning before any text is touched.
9. Stanford HCI research found that writers felt most ownership over drafts when AI was used to:
Correct. Options mode — alternatives rather than replacements — preserved writer agency and produced significantly higher authenticity ratings.
The Stanford finding: asking AI for options (several alternatives to a specific sentence) rather than full rewrites produced highest authenticity and satisfaction ratings.
10. The New York Times 2023 AI guidelines required journalists to:
Correct. The Times required transformation — replacing AI's generic framing with specific reporter knowledge — not just verification or polish.
The Times guidelines required substantial rewriting through reporter judgment — not disclosure alone or research-only restrictions.
11. Which of the four injection points is described as "highest-leverage"?
Correct. AI almost always opens generically — making a specific, surprising personal opener the single highest-impact swap available.
The opening sentence is highest-leverage because AI defaults to generic contextual openers, making any specific personal replacement immediately differentiating.
12. The 40% rule suggests that below 40% writer-originated sentences:
Correct. Below 40% original or substantially transformed sentences, AI's statistical center tends to win — leaving the writer's name on the piece but not their voice in it.
The 40% threshold is a quality benchmark: below it, AI's generic framing tends to dominate and the writer's specific perspective loses influence over the reader's experience.
13. AI examples in drafts are described as "the most generic things" because:
Correct. AI examples are the first-search-result illustrations of a concept — the most commonly cited, therefore the least distinctive.
AI examples are the most common illustrations of any concept — first-search-result-level. Replacing them with examples from your own experience and reading is one of the strongest voice-injecting moves available.
14. "Sentence-level ownership" is best described as:
Correct. Sentence-level ownership is about understanding and authority — not about who typed the words, but whether you can defend each sentence from your own knowledge base.
Sentence-level ownership is the habit of not publishing anything you couldn't explain from your own knowledge — a standard of authority, not of who physically wrote the words.
15. The module's culminating principle — "AI builds the scaffolding; you have to live in it" — means:
Correct. The scaffolding shows where the rooms are; only a writer who has genuinely inhabited the subject can make readers feel what it's like to be there.
The scaffolding metaphor means AI can provide structure, but the felt quality of voice — making readers feel they're in the presence of someone who truly knows the subject — requires genuine inhabitation that no AI generation can substitute.