L1
·
Quiz
·
Lab
L2
·
Quiz
·
Lab
L3
·
Quiz
·
Lab
L4
·
Quiz
·
Lab
Module Test
Module 3 · Lesson 1

How AI Generates Text — and Why It Sounds Like Everyone

Understanding the statistical engine behind AI prose is the first step to not becoming it.
If AI is trained on the entire internet, whose voice does it default to?

In April 2023, the editors of Science journal flagged an unusual pattern: peer-review reports submitted by multiple independent reviewers contained nearly identical phrasing — phrases like "commendable" and "meticulous" appearing in clusters. Researchers at Stanford traced the pattern to ChatGPT. Reviewers under deadline pressure had used the model to draft feedback, and the model had, predictably, converged on its statistical favorites. The result was not plagiarism in any traditional sense — each reviewer had typed different prompts — but the output voices had collapsed into one.

The Stanford study, published in PLOS ONE in January 2024, estimated that between 6% and 16.9% of peer-review text submitted in late 2023 had likely been AI-generated. The signal was unmistakable precisely because the model's vocabulary peaks are so narrow.

What Large Language Models Actually Do

A large language model (LLM) does not "write." It predicts the next most statistically likely token — a word fragment — given everything that came before it. That training corpus is enormous: Common Crawl, Wikipedia, GitHub, books, Reddit, news archives. But enormous does not mean diverse in the way human culture is diverse. The web over-represents English, over-represents certain registers (formal, mildly journalistic, listicle-style), and over-represents the kind of competent-but-forgettable prose that gets published in volume.

The consequence is what researchers call distributional homogenization. Ask ten people to independently describe a sunset and you will get ten genuinely different sentences. Ask ChatGPT, Claude, or Gemini to do the same and, across runs and across models, the outputs cluster. Words like "tapestry," "delve," "testament," "showcasing," and "nuanced" appear at rates far above their frequency in human writing because they sit near the peaks of the model's probability distribution for formal descriptive prose.

Temperature — the randomness parameter — can spread the distribution, producing wilder word choices. But a high-temperature output is random, not personal. Randomness is not a voice.

Why This Matters for Creative Writers

If you use AI output directly, you are not expressing your voice — you are publishing the average of the internet's voice. The goal of this module is to make AI a drafting tool that you edit back into your own idiom, not a ghostwriter that replaces it.

The Vocabulary Gap

Ethan Mollick, a Wharton professor who studies AI adoption, ran an informal experiment in 2023 documented in his Substack "One Useful Thing": he asked GPT-4 to write in twenty distinct author styles. Across all twenty, certain grammatical constructions appeared regardless of style — compound sentences joined by semicolons in a very specific rhythm, a preference for the em-dash as a pause, and a resolution toward optimism at paragraph's end. Style prompting shifted surface vocabulary but rarely the underlying cadence.

This is the vocabulary gap: the distance between the statistical center of mass that AI defaults to and the specific, idiosyncratic vocabulary that makes a writer recognizable. Cormac McCarthy's refusal to use commas, Joan Didion's fragmentary syntax, David Foster Wallace's footnote-as-thought — none of these are behaviors a prompted LLM sustains under its own momentum for more than a few sentences before reverting to the mean.

Key Concepts
Token predictionThe fundamental mechanism of LLMs: each output word is the statistically most likely continuation of the input. Style is a statistical bias, not a learned intention.
Distributional homogenizationThe tendency of AI outputs to converge on the same vocabulary and syntax regardless of who prompted them or what style was requested.
TemperatureA parameter (0–1+) controlling output randomness. Higher values produce more varied but less coherent text; lower values produce more predictable, polished-sounding prose.
VoiceIn writing, the consistent set of word choices, rhythms, syntactic habits, and tonal attitudes that make a writer's work recognizable across pieces and time.
The Practical Rule

Think of AI-generated prose as a rough clay shape — roughly right in form, but not yet carrying your fingerprints. Your job as the writer is to press those fingerprints in at every level: word choice, sentence rhythm, what you leave out, and the specific angle from which you see the world.

Lesson 1 Quiz

How AI Generates Text — and Why It Sounds Like Everyone
1. What did the Stanford / PLOS ONE 2024 study reveal about peer-review text submitted in late 2023?
Correct. The PLOS ONE study estimated 6–16.9% of reviewed text was AI-generated, identified by vocabulary clustering around words like "commendable" and "meticulous."
Not quite. The key finding was vocabulary clustering from AI use — reviewers used AI independently and the outputs converged, not because of copying.
2. At the most fundamental level, what does an LLM do when it generates text?
Correct. Token-by-token statistical prediction is the core mechanism — not retrieval, not rule application, not curation.
Incorrect. LLMs are fundamentally next-token predictors. They don't retrieve passages or apply explicit rules.
3. What does "distributional homogenization" mean in the context of AI writing?
Correct. Distributional homogenization is the convergence of outputs toward statistical peaks regardless of different prompts or users.
Not correct. Homogenization refers to outputs collapsing toward common vocabulary peaks, not to training data processing.
4. Ethan Mollick's style-prompting experiment showed that GPT-4:
Correct. Style prompts moved surface features but not deep structural habits like semicolon rhythm and paragraph-ending optimism.
Incorrect. Mollick found persistent cadence patterns — em-dash pauses, semicolon rhythms — across all twenty style conditions.
5. Why is a high-temperature AI output still not the same as a writer's voice?
Correct. Voice emerges from consistent personal choices made with intention. Random variation mimics neither consistency nor intention.
Not quite. Temperature controls distributional spread; high settings increase variety but randomly, which is fundamentally different from the consistent, intentional choices that constitute a voice.

Lab 1: Voice Fingerprinting

Identify the statistical signatures in AI prose — then locate your own alternatives.

What You'll Practice

Paste a short AI-generated paragraph into the chat and ask the assistant to identify any "distributional homogenization" markers — words or structures that appear at AI-frequency peaks. Then ask it to suggest replacements drawn from a style or vocabulary you actually use. This lab trains the editorial eye you need to reclaim your voice from AI drafts.

Try starting with: "Here is a paragraph I drafted with AI help: [paste 3–5 sentences]. What vocabulary or cadence markers suggest it defaults to AI statistical averages? What might replace them if my writing tends toward [describe your style]?"
Voice Fingerprinting Assistant
Lab 1
Welcome to Lab 1. Paste an AI-generated paragraph — or describe the kind of prose you want to analyze — and I'll help you identify statistical-average markers and suggest replacements that fit your actual voice. Where would you like to start?
Module 3 · Lesson 2

Prompting for Draft Material, Not Finished Prose

The professional workflow: AI as quarry, not architect.
What changes when you treat AI output as raw material instead of a deliverable?

In September 2022, novelist Robin Sloan — author of Mr. Penumbra's 24-Hour Bookstore — published a detailed essay in the Atlantic describing how he used GPT-3 during the writing of his novel Moonbound. Sloan's method was specific: he fed the model a scene description and asked it to generate ten possible sentences that a character might say. He used none of them verbatim. He used them as "a tuning fork," his phrase — something to push against. If the model produced a flat declarative, he wrote a run-on. If it produced optimism, he reached for irony. The AI was not a co-author; it was a negative space he wrote into.

Sloan was explicit about what made this work: he had already written two novels. He had enough command of his own idiom that he could feel immediately when the AI was pulling him toward the generic. Writers without that established self-knowledge, he warned, risk "learning the AI's voice instead of their own."

The Draft-Material Mindset

The single most important conceptual shift for creative writers using AI is this: never treat AI output as a candidate for publication. Treat it as a quarry — a source of raw stone from which you cut your actual work. This is not a moral position about AI authorship; it is a practical one about voice and quality.

When you submit an AI draft directly, you are inheriting all its statistical defaults: its vocabulary peaks, its preferred sentence rhythms, its tendency to resolve ambiguity toward clarity, its aversion to the kinds of syntactic risk that distinguish memorable prose. When you use it as quarry, you extract what is useful (a structural skeleton, a list of possible metaphors, a factual summary to paraphrase) and discard the rest.

The workflow distinction looks like this: Submitting AI drafts → you become an editor of someone else's voice. Using AI for material → you remain the writer, with a richer palette of raw options.

The Washington Post Approach (2023)

The Washington Post's technology team documented their internal AI writing guidelines in October 2023. Reporters were permitted to use AI to generate initial research summaries and to draft structural outlines, but all prose submitted for publication had to be written by the reporter. The guideline explicitly stated: "AI-generated sentences should not appear in published text." The rationale was not just accuracy — it was voice. Editors had noticed that AI-drafted ledes flattened the tonal signature that distinguished Post reporting from wire copy.

Prompt Architectures for Draft Material

The prompts that generate useful raw material differ structurally from prompts that generate finished prose. Useful material prompts tend to:

1. Request plurality. "Give me twelve possible opening lines for this essay" produces a range from which you choose and modify. "Write the opening line" produces one answer you either use or reject wholesale.

2. Request raw ingredients, not assembled dishes. "List ten concrete images that could represent isolation in a city" gives you material. "Describe isolation in a city" gives you prose you will edit forever without making it yours.

3. Set explicit constraints that force departure from defaults. "Draft a paragraph about grief that uses no abstract nouns and no metaphors" pushes the model away from its statistical comfort zone and produces something stranger and more usable.

4. Request structural scaffolding, not prose. "Give me a three-section outline for this argument, with one counterargument per section, as bullets only" is infinitely more useful than asking for a draft essay, because you will write the sentences yourself.

Key Terms
Draft-material mindsetTreating AI output as a quarry of raw options — images, structures, alternatives — rather than as candidate finished text.
Plurality promptingAsking AI to generate multiple options (ten openings, twelve images) rather than a single answer, preserving your choice and editorial control.
Constraint promptingSpecifying unusual structural constraints (no abstract nouns, only present tense, exactly seven words per sentence) to push AI output away from statistical defaults.
Sloan's Tuning Fork Principle

The most experienced writers use AI output as something to write against, not from. If the AI produces the obvious, write the non-obvious. If it resolves toward hope, consider whether your piece needs ambiguity. The model's defaults can clarify your own choices by opposition.

Lesson 2 Quiz

Prompting for Draft Material, Not Finished Prose
1. How did Robin Sloan use GPT-3 during the writing of Moonbound, according to his 2022 Atlantic essay?
Correct. Sloan explicitly described using AI output as a tuning fork to define his own choices by opposition, never using generated sentences verbatim.
Incorrect. Sloan was careful to use no verbatim AI text. The model was a pressure to write against, not a source to pull from.
2. What was the Washington Post's documented AI guideline for reporters in October 2023?
Correct. The Post allowed AI for structural and research purposes but prohibited AI-generated sentences in published text, citing voice flattening as the concern.
Not correct. The Post permitted AI for research summaries and outlines — the restriction was specifically on AI-generated prose appearing in published articles.
3. What is "plurality prompting" and why does it preserve a writer's voice?
Correct. Plurality prompting restores editorial control: you select from options, which keeps you as the decision-maker and prevents any one AI output from dominating.
Incorrect. Plurality prompting means asking for many alternatives at once — ten lines, twelve images — so the choice of what to use and how to modify it remains entirely with the writer.
4. Why does asking AI for "concrete images" rather than "descriptive prose" produce more useful raw material?
Correct. When the writer assembles raw ingredients into sentences, that assembly process is where personal rhythm, word choice, and perspective enter the work.
Not correct. The key distinction is that prose prompts produce an assembled AI voice; ingredient prompts give you components to assemble yourself — and the assembly is where your voice lives.
5. Sloan warned that writers without an established voice risk which specific problem when using AI?
Correct. Sloan's exact warning: writers without established self-knowledge risk "learning the AI's voice instead of their own" — internalizing AI defaults as if they were personal preferences.
Not quite. Sloan's warning was specifically that less experienced writers might unconsciously absorb AI voice patterns, confusing statistical defaults with their own developing style.

Lab 2: Plurality and Constraint Prompting

Practice generating material you can actually use — without surrendering authorship.

What You'll Practice

In this lab you'll work with the assistant to build plurality prompts and constraint prompts for a writing project you have in mind. The goal is to produce raw material — not finished prose — and then identify which elements you would actually use and why.

Try: "I'm working on [describe your project in one sentence]. Help me write a plurality prompt that will give me ten concrete options for [a specific element], and a constraint prompt that pushes the AI away from generic phrasing."
Draft Material Lab
Lab 2
Welcome to Lab 2. Tell me about a writing project you're working on — even just a topic or genre — and we'll build plurality prompts and constraint prompts together that generate useful raw material without handing your voice to the AI. What are you working on?
Module 3 · Lesson 3

Editing AI Text Back Into Your Voice

The revision techniques that reclaim authorship from a statistical machine.
What specific edits transform AI prose into your prose?

In December 2022, Stephen Marche — author and cultural critic — published an essay in The Atlantic with an unusual credit line: the piece had been written collaboratively with GPT-3, Sudowrite, and Cohere. Marche documented the process in detail. He found that AI could produce competent sentences quickly but that those sentences had a specific defect: they resolved. They concluded. They moved toward the summarizing gesture. His own writing style — recursive, looping back on itself, comfortable with unresolved tension — had to be actively imposed on every paragraph through revision.

Marche's editing method was systematic: he identified every sentence that used an abstract noun where a concrete one would serve, every paragraph that ended with a claim instead of an image, every transition word ("therefore," "moreover," "additionally") that appeared in the AI draft and deleted it. Then he rewrote each of those positions from scratch. The result, he said, read like him. The process took longer than writing from scratch — but produced a richer draft to push against.

The Five-Layer Edit

Editing AI text back into your voice requires working at five distinct levels simultaneously, because AI defaults operate at all five:

1. Vocabulary. Replace the model's statistical peak words with your actual vocabulary. This means having a clear enough sense of your own word preferences to know that you would never write "delve," that you prefer "look into" or "dig through" or sometimes just "read." If you don't yet know your vocabulary preferences well, read ten pages of your best previous writing and list the words that appear there but not in AI drafts.

2. Sentence rhythm. Count the syllables in three consecutive AI sentences. They will often be close to each other — the model prefers rhythmic regularity. Your own prose likely has more variation: a short sentence. Then a longer one that accumulates clauses in a specific way. Then another short one for emphasis. Rewrite to match your rhythm, not the model's.

3. Resolution habit. AI defaults toward closure and clarity. If your writing is more comfortable with ambiguity, look at every paragraph ending. If it ends with a declaration, consider replacing it with a question, an image, or a withheld conclusion.

4. Transition vocabulary. "Furthermore," "moreover," "additionally," "in conclusion," "it is worth noting" — these are AI transition staples. Delete all of them and either write your own connection or, if the logic is obvious, cut the transition entirely. Your own transition vocabulary is more specific and probably more idiosyncratic.

5. Specificity level. AI tends toward the category level: "a bird," "a building," "an emotion." Your prose lives at the instance level: "a red-tailed hawk," "a pre-war tenement on Myrtle Avenue," "the specific anxiety of someone about to speak in front of a crowd they have underestimated." Go through the AI draft and push every abstraction toward a specific instance.

The Turnitin Observation — 2023

Turnitin's AI detection research team, in a white paper released in August 2023, documented that the most commonly flagged AI prose patterns were not rare vocabulary items but structural signatures: paragraphs that averaged similar lengths, consistent use of three-part sentence structure, and transition word density above the 90th percentile of human academic writing. Editors at major magazines reported independently that AI copy "felt smooth in a way real writing rarely is" — meaning the absence of idiosyncratic rough edges was itself the tell.

The "Rough Edge" Principle

The Turnitin finding points to something important: your voice is partly constituted by what a statistical model would never generate — the specific imperfections, departures from convention, and personal tics that feel wrong by average standards but right by the standards of your particular sensibility.

David Sedaris's run-on sentences that break grammatical rules. Zadie Smith's intrusive narrator who interrupts the fiction to comment on it. Joan Didion's refusal to provide context before emotion. These are rough edges. They are what editing AI prose toward smoothness removes, and what editing AI prose toward your voice requires you to add back in.

A practical exercise: take a paragraph you have written in the past that you consider among your best. Count the "grammatical violations" in it — comma splices, fragments, unconventional punctuation. Now look at the AI draft you are editing. Count the same. If the AI draft has fewer violations, it is too smooth. Add some of yours back in.

Key Terms
Five-layer editA systematic revision process addressing vocabulary, sentence rhythm, resolution habit, transition vocabulary, and specificity level to reclaim AI prose for a human voice.
Resolution habitAI's statistical tendency to end paragraphs with declarative conclusions rather than images, questions, or ambiguity.
Rough edge principleThe concept that a writer's voice is partly constituted by intentional departures from convention — the specific imperfections that statistical averaging eliminates.

Lesson 3 Quiz

Editing AI Text Back Into Your Voice
1. What specific defect did Stephen Marche identify in all AI-generated prose during his 2022 Atlantic collaboration?
Correct. Marche's core critique: AI prose resolves. His own writing loops, defers conclusion, and lives with unresolved tension — the opposite of the AI default.
Incorrect. Marche's specific observation was about resolution and conclusion — AI paragraphs moved toward summarizing gestures when his style is recursive and resistant to closure.
2. According to Turnitin's August 2023 white paper, what were the most reliably flagged AI prose patterns?
Correct. Turnitin found structural uniformity — not vocabulary — was the strongest AI signal: paragraph length consistency, three-part sentences, and transition word overuse.
Not correct. Turnitin's research showed structural signatures were the strongest flag: uniformity of paragraph lengths, predictable sentence structure, and high-density transition words.
3. The "rough edge principle" states that a writer's voice is partly constituted by:
Correct. Voice is not conformity — it is a specific pattern of departures from the norm that a reader comes to recognize and expect from that writer.
Incorrect. The rough edge principle is specifically about intentional violations of convention — comma splices, fragments, unconventional punctuation — that constitute recognizable voice.
4. In the five-layer edit, what does addressing "specificity level" require?
Correct. AI operates at category level; human voice operates at instance level. Specificity is one of the clearest markers of genuine observation and personal experience.
Not correct. Specificity editing means pushing from the AI's generic categories toward the specific instances — particular birds, particular streets, particular feelings — that make writing feel lived-in.
5. Marche's editing of his AI-assisted Atlantic piece took longer than writing from scratch. Why did he consider the process worthwhile?
Correct. The value was in having a rich, resistible draft — something substantial enough to push against — not in saving time. The final product was his voice, achieved through more revision, not less.
Incorrect. Marche explicitly noted the process took longer, but the AI draft gave him richer material to shape. Time was not the benefit — quality of the raw material was.

Lab 3: The Five-Layer Edit

Practice the systematic revision that moves AI prose into your voice.

What You'll Practice

Bring an AI-generated paragraph (or ask the assistant to generate one for you to practice on). Together you'll work through the five-layer edit: vocabulary, sentence rhythm, resolution habit, transition vocabulary, and specificity. The assistant will identify which layers need work and suggest revision strategies — but you make the actual word choices.

Try: "Please generate a generic three-sentence paragraph about [any topic]. Then walk me through the five-layer edit, identifying what would need to change to match a writer whose style is [describe briefly]."
Five-Layer Edit Assistant
Lab 3
Welcome to Lab 3. Paste an AI-generated paragraph, or ask me to generate one, and we'll run it through the five-layer edit together: vocabulary, sentence rhythm, resolution habit, transition words, and specificity. I'll flag what needs changing — but you'll do the rewriting. Ready to start?
Module 3 · Lesson 4

Disclosure, Attribution, and the Ethics of AI-Assisted Writing

What the emerging norms require — and why transparency protects the writer.
When does using AI become something you need to disclose, and to whom?

In January 2023, Clarkesworld — one of the most prestigious science fiction magazines — closed its submissions window after receiving a flood of AI-generated stories. Editor Neil Clarke documented the numbers publicly: in the three months following ChatGPT's public release, AI-generated submission attempts had increased by approximately 800 percent. Clarke could identify most of them by style markers, but not all. He reopened submissions with a new policy: any story suspected of AI generation would be rejected without feedback, and submitters found to have disguised AI authorship would be permanently blacklisted.

Within the same month, three other SFWA (Science Fiction and Fantasy Writers of America) affiliated publications — Beneath Ceaseless Skies, Strange Horizons, and Fantasy & Science Fiction — posted explicit AI prohibitions. The prohibitions were not primarily about copyright; they were about voice. As Strange Horizons stated: "We publish writers, not outputs."

The Disclosure Landscape in 2024

Disclosure norms for AI-assisted writing are evolving rapidly and vary significantly by context. Understanding what each context requires is now a basic professional competency for any working writer.

Academic writing: The Modern Language Association (MLA) and the American Psychological Association (APA) both issued guidance in 2023 requiring disclosure of AI tools used in the preparation of manuscripts. The MLA stated that AI-generated text should be treated as a source requiring citation. The APA required that any use of AI tools be described in the methods or acknowledgements section. Many individual universities have gone further, treating undisclosed AI use as academic dishonesty equivalent to plagiarism.

Journalism: The Society of Professional Journalists (SPJ) added AI guidance to its Code of Ethics in 2023: reporters must disclose AI use to editors, and publications must disclose AI-generated or AI-substantially-assisted content to readers. The Associated Press style guide now includes entries on AI attribution.

Commercial and creative writing: Publishing contracts signed from mid-2023 onward have increasingly included "AI warranty" clauses in which the author certifies that the work is not substantially AI-generated. The Authors Guild's model contract language specifies that the author must disclose if AI tools were used in the composition of more than a de minimis portion of the work.

The De Minimis Question

The phrase "de minimis" — meaning "too small to matter" — appears in multiple emerging AI disclosure frameworks, but no consensus definition exists yet. Using AI to suggest a synonym is almost universally considered de minimis. Using AI to draft a full section and editing it is not. The practical guidance: when in doubt, disclose. A brief acknowledgement ("AI tools were used in drafting this piece") costs nothing and protects against accusations of bad faith.

Copyright: What You Do and Don't Own

In February 2023, the U.S. Copyright Office issued a guidance statement on AI-generated works: content generated entirely by AI is not copyrightable. Copyright requires human authorship. In August 2023, the Office clarified for the graphic novel case Zarya of the Dawn (Kris Kashtanova): the AI-generated images were not copyrightable, but the human-authored text and arrangement were. This is the current legal framework in the U.S.: your creative decisions are copyrightable; the AI's text generation is not.

The practical implication for writers: the more you edit, restructure, and rewrite AI output, the stronger your copyright claim to the result. A lightly edited AI draft may have ambiguous copyright status. A heavily revised AI draft that retains few of the original AI sentences is substantially your own work. This is another reason — beyond voice — that the five-layer edit matters.

Why Transparency Protects Writers

The instinct to conceal AI use is understandable but strategically counterproductive. Editors, academic reviewers, and readers are becoming more capable of detecting AI-influenced prose. Being discovered to have concealed AI use after the fact causes far more reputational damage than proactive disclosure. The writers who have navigated this period most successfully — Marche at The Atlantic, Sloan with his documented process, journalists at publications that have developed explicit AI policies — are those who have been transparent about their workflow and specific about what AI did and did not contribute.

Key Terms
De minimis useAI assistance so limited (a synonym suggestion, a spell-check) that disclosure is generally not required by emerging norms. No universal threshold has been agreed upon.
AI warranty clauseA contractual provision, increasingly standard in publishing contracts from 2023 onward, in which the author certifies the degree of AI use in the work's composition.
Human authorship requirementThe U.S. Copyright Office's position that copyright requires a human author; AI-generated content without substantial human creative input is not copyrightable.
The Practical Rule for All Contexts

Before submitting any AI-assisted work, ask three questions: (1) Does the publication or institution have an explicit AI policy? If yes, follow it. (2) Have I substantially transformed the AI output through revision? If yes, the work is primarily yours. (3) Would a reasonable reader or editor consider my disclosure sufficient? If uncertain, add more disclosure. Transparency is never the wrong choice.

Lesson 4 Quiz

Disclosure, Attribution, and the Ethics of AI-Assisted Writing
1. What happened at Clarkesworld magazine in January 2023, and what policy change resulted?
Correct. Editor Neil Clarke documented the ~800% surge and responded by closing submissions temporarily and implementing a blacklist for disguised AI authorship upon reopening.
Incorrect. The crisis was an 800% surge in AI-generated story submissions that overwhelmed the editorial process. The response was temporary closure and a strict new policy against undisclosed AI authorship.
2. What did the U.S. Copyright Office rule in February 2023 regarding AI-generated content?
Correct. The Copyright Office stated clearly: copyright requires a human author. AI-generated text, without substantial human creative input, is not protectable.
Not correct. The Copyright Office ruled that AI-generated content lacks the human authorship copyright law requires. The AI company holds no copyright; neither does the prompter for purely AI-generated text.
3. The MLA's 2023 AI guidance stated that AI-generated text should be treated as:
Correct. The MLA treated AI-generated text as a source to be cited — requiring both citation of the tool and disclosure of how it was used.
Incorrect. The MLA's position is that AI-generated text is a citable source: the tool must be identified and its use disclosed, similar to how any other source is cited.
4. What practical copyright implication does the Zarya of the Dawn ruling have for writers who heavily edit AI drafts?
Correct. Zarya clarified that human creative decisions (arrangement, text) are copyrightable even in AI-assisted works. More substantial human transformation means stronger copyright protection.
Incorrect. The Zarya ruling specifically protects human creative decisions within AI-assisted works. Substantial revision strengthens the copyright claim because the creative decisions are then clearly the human author's.
5. What is an "AI warranty clause" in a publishing contract?
Correct. AI warranty clauses, increasingly common in contracts from mid-2023, require authors to certify the degree of AI involvement in their manuscript's composition.
Not correct. An AI warranty clause is a certification by the author about how much AI was used in writing the work — it protects the publisher from undisclosed AI authorship.

Lab 4: Disclosure Decisions in Practice

Work through real disclosure scenarios for your writing context.

What You'll Practice

Disclosure norms vary by context, and the right decision is not always obvious. In this lab, describe a specific writing situation — academic paper, magazine pitch, fiction submission, commercial copy, blog post — and the degree of AI involvement you used. The assistant will help you determine what disclosure is required, recommended, and sufficient given current guidelines from the relevant professional bodies.

Try: "I used AI to generate an outline and three body paragraphs, which I then rewrote substantially. I'm submitting this as a [type of piece] to [type of publication or context]. What disclosure do I need to make, if any, and what language should I use?"
Disclosure Advisor
Lab 4
Welcome to Lab 4. Describe your situation: what kind of piece are you writing, where are you submitting it, and how did you use AI in the process? I'll help you work through what disclosure is required, what's recommended, and what language would be appropriate. What's your scenario?

Module 3 Test

Writing, Editing, and Staying Your Own Voice — 15 questions, 80% to pass
1. The PLOS ONE 2024 study estimated that what percentage of peer-review text submitted in late 2023 showed AI-generation markers?
Correct. The Stanford/PLOS ONE study estimated 6–16.9% of peer-review text had likely been AI-generated.
The study found 6–16.9%, detected by vocabulary clustering around words AI models statistically favor.
2. What is "token prediction" in the context of LLM text generation?
Correct. Token prediction is the core mechanism: each word-fragment is generated as the statistically most probable continuation.
Token prediction means each token (word fragment) is chosen based on statistical probability given what came before — not measurement, tracking, or cryptography.
3. In Robin Sloan's documented workflow for Moonbound, what was the AI's role?
Correct. Sloan used AI output as a pressure to write against — defining his choices by opposition rather than by adoption.
Sloan's method was the opposite of adoption: AI produced sentences he wrote against, using none verbatim — a "tuning fork" not a drafter.
4. Which of the following best describes "plurality prompting"?
Correct. Plurality prompting requests many options at once, restoring the writer's role as selector and editor rather than acceptor of a single AI output.
Plurality prompting means requesting multiple alternatives from a single prompt — ten lines, twelve images — so the choice and modification remain with the writer.
5. What is the "draft-material mindset" as applied to AI writing assistance?
Correct. The draft-material mindset treats AI output as stone to cut from, not as architecture to inhabit.
The draft-material mindset is about treating AI output as quarry — raw material to extract useful elements from — while the writer provides the actual prose.
6. Stephen Marche's primary editorial technique when revising his AI-assisted Atlantic piece was to:
Correct. Marche systematically deleted abstract nouns, conclusion-paragraphs, and transition words, then rewrote each deleted position in his own idiom.
Marche's method was surgical deletion — abstractions, closing declarations, and transition words — followed by rewriting those exact positions from scratch.
7. Turnitin's 2023 white paper found that AI prose was most reliably detected by:
Correct. Structural uniformity — not unusual vocabulary — was Turnitin's strongest AI detection signal.
Turnitin found structural signatures were the tell: uniform paragraph lengths, predictable three-part sentence structure, and transition word density above the 90th percentile.
8. In the five-layer edit, "resolution habit" refers to:
Correct. AI defaults toward closure and clarity at paragraph endings; writers who prefer ambiguity or image-based endings must actively override this habit.
Resolution habit is about paragraph-ending behavior: AI consistently moves toward declarative summary; writers who live with ambiguity must intervene at every paragraph end.
9. The "rough edge principle" holds that a writer's voice is partly constituted by:
Correct. Recognizable voice often lives in its rule violations — comma splices, fragments, unconventional rhythm — all of which AI averaging removes.
The rough edge principle is the opposite of conformity: intentional grammatical violations and syntactic idiosyncrasies are what make a voice recognizable across time.
10. What was Clarkesworld's policy response to the 2023 AI submission surge?
Correct. Neil Clarke closed submissions, documented the surge publicly, and implemented a permanent blacklist for disguised AI authorship upon reopening.
Clarkesworld closed temporarily due to an ~800% surge in AI submissions, then reopened with a blacklist policy for anyone found to have disguised AI authorship.
11. The U.S. Copyright Office's position on AI-generated content, established in 2023, is:
Correct. The Copyright Office established clearly: copyright requires human authorship. Purely AI-generated text is not protectable.
The Copyright Office ruled that AI-generated content — lacking human authorship — is not copyrightable. The AI company holds no copyright; neither does the prompter for purely generated text.
12. The MLA's 2023 guidance treats AI-generated text as:
Correct. The MLA treats AI tools as citable sources requiring disclosure of the tool and the manner of its use — similar to citing any secondary source.
The MLA requires citation: the AI tool must be named and the manner of use disclosed, treating it as a source rather than an invisible tool.
13. "Constraint prompting" helps writers avoid AI statistical defaults by:
Correct. Unusual constraints push AI away from its statistical peaks, producing stranger, less generic output that requires less editorial effort to reclaim.
Constraint prompting means imposing structural rules that AI defaults wouldn't produce — no abstract nouns, only fragments, seven-word sentences — pushing it toward less generic territory.
14. The Zarya of the Dawn copyright ruling (August 2023) established which principle relevant to AI-assisted writers?
Correct. Zarya: the human-authored text and creative arrangement were protectable; the AI-generated images were not. Human decisions, not AI outputs, qualify for copyright.
Zarya established that human creative decisions within an AI-assisted work are protectable. AI outputs in the same work are not — but the human's choices of text, structure, and arrangement are.
15. Which of the following best summarizes why transparency about AI use protects writers professionally?
Correct. Editors and reviewers are increasingly capable of detecting AI-influenced prose. Proactive disclosure costs little; retroactive discovery after concealment costs enormously.
The practical case for transparency: AI detection is improving, and being discovered to have concealed use is far more damaging than disclosing upfront. Transparency is the professionally safe choice.