Intro
L1
·
Quiz
·
Lab
L2
·
Quiz
·
Lab
L3
·
Quiz
·
Lab
L4
·
Quiz
·
Lab
Module Test
Make It Yours: Create With AI · Introduction

Every Tool That Changed Making Also Changed Who Got to Make

Why this course exists — and what you'll actually be able to do when it's done.

In 1888, George Eastman put a camera called the Kodak on sale for twenty-five dollars and changed the meaning of the word "photographer." Before that moment, making a photograph required a studio, chemicals, glass plates, and years of technical apprenticeship. After it, anyone who could press a button could capture an image. Professional photographers warned that the craft would be devalued, that the masses would produce nothing but mediocrity. They were partly right and mostly wrong: the overall volume of creative output exploded, new genres emerged that the professionals had never imagined, and a century later the photographic tradition is considered richer, not poorer, for the democratization.

Something structurally similar is happening right now with language, code, music, and image. In November 2022, OpenAI released ChatGPT to the public; within two months it had a hundred million users — the fastest consumer-product adoption ever recorded. By mid-2023, Adobe had embedded generative AI into Photoshop, Spotify was hosting AI-voiced podcast translations, and publishers were debating what "authorship" even meant. The tools are not finished, the legal frameworks are not settled, and the creative questions are genuinely open.

This course is a working introduction to AI as a creative collaborator. You will learn how to write prompts that actually produce what you intend, how to iterate when the first result is wrong, how to use AI for visual and musical ideas, and how to think clearly about what remains yours when a machine helped. It will not make every creative problem easy. It will give you a real starting vocabulary and a set of practiced instincts — so that whatever these tools become next, you are not starting from zero.

Make It Yours: Create With AI · Lesson 1

The Blank Page Just Got a Partner

What AI creative tools actually are, how they work at a useful level, and why the prompt is the most important thing you will write.
What changes — and what doesn't — when the hardest part of creating is no longer starting?

In January 2023, the novelist Kazuo Ishiguro was asked in a BBC interview whether AI could write a novel. He said he doubted it — not because the sentences would be bad, but because a novel is made of things the author genuinely does not know yet, things that surprise the writer during the writing. The machine, he argued, does not not-know anything. It produces the statistically expected next word. That is a different activity than discovery.

Ishiguro is probably right about the deepest layer. But there is a large territory between "producing statistically expected words" and "writing a novel as an act of self-discovery" — and that territory is where most actual creative work happens. Drafting. Generating options. Breaking blocks. Finding a phrase you then improve. This is where AI tools are genuinely useful right now, and this is where we will spend most of our time.

The critical insight is that the quality of what comes out is tightly coupled to the quality of what goes in. A vague prompt produces vague output. A specific, contextually rich prompt produces something you can actually use — or at least usefully react to. Learning to write that prompt is the skill this course is built around.

What "Generative AI" Actually Means

Large language models — GPT-4, Claude, Gemini, Llama — were trained on enormous text datasets by a process that, simplified, involves predicting which token comes next in a sequence. After enough of that training, the model develops internal representations that allow it to generalise: it can complete sentences it has never seen, adopt registers it was not explicitly taught, and reason about topics by analogy to patterns it absorbed. This is not the same as understanding in the human sense, but it is also not simple autocomplete.

Image generators like Midjourney, DALL·E 3, and Stable Diffusion work differently: they learn to map between text descriptions and image features through a process called diffusion — starting from noise and iteratively removing it according to what the prompt implies. Music generators like Suno (launched publicly in 2023) and Udio use related architectures applied to audio tokens.

You do not need to understand the mathematics to use these tools well. But you do need to understand one practical implication: the model has no intention. It does not want to help you write a good story. It is producing the most probable continuation of the context you gave it. Your job, as the human, is to construct a context — a prompt — that makes the probable continuation also the useful one.

The Anatomy of a Prompt

Researchers at Anthropic, OpenAI, and independent prompt-engineering teams have converged on a rough taxonomy of what makes prompts work. The following components are not rules — they are levers. You choose which to pull for a given task.

Role
Who the model should act as. "You are a copywriter who specialises in product launches for B2B software." Setting role shifts register, vocabulary, and default assumptions.
Task
The specific action requested. Not "write something about my project" but "write a 150-word opening paragraph for a Kickstarter campaign page."
Context
Background the model needs but doesn't have: your audience, the tone of your brand, what has already been said, what must not be repeated.
Format
How the output should be structured: bullet list, paragraph, dialogue, table, numbered steps, JSON, a specific word count.
Constraints
What to avoid. "Do not use the word 'innovative.' Do not make claims about ROI. Keep sentences under 20 words."
Examples
One or two samples of the style or output you want. This is called "few-shot prompting" and is often the single highest-value addition to a prompt.
Why This Matters

In 2023, Stanford researchers studying GitHub Copilot (the AI coding assistant used by over a million developers) found that programmers who wrote more specific, contextual prompts accepted AI suggestions at twice the rate of those who wrote brief prompts — and the accepted suggestions required significantly less editing. The pattern holds in writing and design tasks too. Specificity is the primary quality lever.

Iteration Is the Method

Professional prompt engineers do not write one prompt and expect a final result. They treat the first output as a draft — material to react to. The discipline is: run, evaluate, diagnose, refine. If the output is too formal, add "conversational and direct" to the format instruction. If it missed a key point, add that point to the context. If the structure is wrong, show an example of the right structure.

This iterative loop is the closest analogy to how editors work with writers, or how directors work with actors. The AI is not the author. You are. The AI is the person in the room who will try anything you ask without complaint, at high speed. Your judgment about what works — your taste, your knowledge of your audience, your sense of your own voice — is the irreplaceable ingredient.

The practical implication: do not judge AI creative tools by their first output. Judge them by what you can coax from them across three or four iterations. Most beginners quit too early. Most experienced users have a specific diagnostic vocabulary — they know which lever to pull when the output is flat, when it is generic, when it is structurally wrong — and that vocabulary is learnable.

The Core Principle of This Module

You are not learning to operate a machine. You are learning to direct a very fast, very capable collaborator who has no taste, no intention, and no investment in the outcome. The directing skill — specific, contextual, iterative — is what this module is about. Everything else is application.

What AI Cannot Do (Yet, and Perhaps Ever)

Ishiguro's point deserves honesty. AI tools do not have genuine aesthetic preferences. They do not know what it feels like to be bored at 2 a.m. trying to find the right word, which is part of where the right word comes from. They do not have the lived experience that gives creative work its specificity and authority.

More practically: they hallucinate. They produce confident-sounding text that is factually wrong. They average across their training data, which means they trend toward the expected. They do not know what happened yesterday. Their outputs on contested cultural or personal topics can be generic to the point of uselessness.

Knowing these limits is not pessimism — it is calibration. A surgeon who knows what a scalpel cannot do is a better surgeon. The limits define where your judgment is not optional: verifying claims, bringing specificity from your own knowledge, and making the final call about whether something is actually good. These remain entirely human responsibilities.

Lesson 1 Quiz

Four questions · Select the best answer for each
1. According to the lesson, what is the single most important factor in the quality of AI creative output?
Correct. The lesson argues that specificity is "the primary quality lever" — vague prompts produce vague output, and specific contextually rich prompts produce something usable or at minimum useful to react to.
Not quite. The lesson's central argument is that prompt quality — specifically the specificity and context you bring — determines output quality more than any technical factor about the model itself.
2. Which of the following is NOT listed as a component in the anatomy of a prompt?
Correct. The lesson lists Role, Task, Context, Format, Constraints, and Examples as prompt components. "Emotional tone score" is not a category — though tone itself can be addressed under Format or Constraints.
Check the key terms section again. Role, Constraints, and Examples are all explicitly listed. The one that does not appear is "emotional tone score" — though tone can be addressed within other components.
3. What does "few-shot prompting" mean in the context of the lesson?
Correct. The lesson defines few-shot prompting under the "Examples" component: "One or two samples of the style or output you want" — and describes it as "often the single highest-value addition to a prompt."
The lesson specifically defines few-shot prompting as including one or two examples of the style or output you want inside the prompt itself — and calls it often the highest-value addition you can make.
4. What was novelist Kazuo Ishiguro's core argument about AI and novel-writing, as described in the lesson?
Correct. Ishiguro's argument was that a novel is made of things the author does not know yet — genuine discovery — and that AI, producing statistically expected continuations, is doing a fundamentally different thing.
Ishiguro's point was more specific than style or grammar. His argument was that the deepest function of novel-writing is discovery — confronting what you genuinely don't know — and that AI, which produces probable continuations, cannot do that.

Lab 1 — Prompt Anatomy in Practice

Build a real prompt from scratch using the six components · Minimum 3 exchanges to complete

Your Task

You are going to write a prompt for a real creative task — then refine it at least twice based on feedback. The AI assistant here knows the six-component prompt framework from Lesson 1 and will help you diagnose why a prompt works or doesn't, suggest which lever to pull, and respond to your drafts as if it were the creative AI you are prompting.

Start by describing a creative thing you actually want to make: a piece of writing, a concept for an image, a title, a product description, anything. Then build your prompt using the framework components. The assistant will both evaluate your prompt structure and show you what the output would look like.

Try starting with: "I want to write a prompt for [describe your creative goal]. Here is my first attempt: [paste your draft prompt]." Then we'll work from there.
AI Prompt Coach
Lab 1
Welcome to Lab 1. Tell me what you want to create — a piece of writing, an image concept, a tagline, anything — and share your first attempt at a prompt for it. I'll help you dissect it using the six-component framework and then show you what a well-structured version produces. What are we making?
Make It Yours: Create With AI · Lesson 2

Writing With AI Without Sounding Like Everyone Else

How to use language models for writing tasks while preserving — and sharpening — your own voice.
If the model was trained on millions of texts, why does everything it writes sound the same — and how do you stop that?

In March 2023, a team at Columbia Journalism Review analyzed a sample of AI-generated press releases submitted to newsrooms. They found a distinctive set of phrases appearing at statistically anomalous rates: "delve," "leverage," "cutting-edge," "in today's rapidly evolving landscape," "it is important to note." These phrases had not been common in human business writing before 2022. They were emerging as AI fingerprints — the model's statistical centre of gravity made visible.

The same pattern shows up in student essays, marketing copy, and social media posts. When a model averages across millions of texts, it tends toward the median — the phrasing that appears in the widest variety of contexts. That median is not bad writing exactly. It is competent, inoffensive, and utterly unmemorable. Your job is to push the output away from the median and toward something that sounds like you knew something specific and cared enough to say it precisely.

Voice Is Information

The reason AI output tends toward blandness is that voice, in writing, is largely made of specific choices: a preference for short sentences or long ones, a habit of starting paragraphs with questions, a tendency to anchor abstractions in particular kinds of concrete detail, a characteristic relationship to humour or irony. These choices are not arbitrary — they are the accumulated trace of a person's reading, thinking, and life. The model doesn't have that trace. It has the average of everyone else's traces.

The practical solution is to treat voice as information you supply in the prompt. This means: giving the model examples of your own writing, describing your preferences explicitly ("I write in short declarative sentences and avoid adverbs"), and using the model's output as a raw material that you then rewrite rather than publish directly.

The rewrite step is not optional if you care about voice. The most effective workflow used by professional writers who use AI — including those at publications like The Atlantic, which discussed AI-assisted workflows openly in 2023 — is to use the model to generate structure and options, then rewrite the passages that will be read closely in their own voice.

Techniques for Voice Preservation

These are not theoretical — they are practiced techniques with documented effectiveness:

Voice samples
Paste 2–3 paragraphs of your own writing into the prompt and say "Match this voice." The model will analyse sentence length, rhythm, vocabulary level, and register and apply those patterns to the new task.
Negative constraints
Explicitly ban the median phrases. "Do not use: delve, leverage, cutting-edge, it is worth noting, in today's landscape, or any similar filler." This single instruction measurably reduces AI fingerprinting.
Temperature and variation
Ask for multiple versions: "Give me three different openings for this paragraph with different tones." Variation gives you options to choose from rather than one statistical centre-of-gravity answer.
Specificity injection
Supply the concrete details yourself. "Write a paragraph about my experience learning to cook, including: my mother's kitchen in Accra, the smell of groundnut soup, a specific Tuesday in 1997." The model can't invent your specifics — but once you supply them, it can build around them effectively.
The rewrite layer
Use AI output as a structural scaffold. Accept the paragraph order and key ideas, then rewrite every sentence in your own words. This is fast — the structural work is done — and the result sounds entirely human because it is.
Real Case

The novelist Gary Shteyngart wrote about using ChatGPT for his 2023 novel "Our Country Friends" promotional work in The New Yorker. He found it useful for generating options he then heavily edited, but noted that the model consistently made his voice "warmer and less strange" — flattening exactly the qualities that make his work distinctive. His solution was to use AI for structural drafts and then aggressively rewrite the surface text. This is now a common professional pattern.

When Generic Is Fine

Not every writing task requires a distinctive voice. Terms and conditions do not need your personality. A first draft of a meeting agenda does not need to sound like you at your literary best. A summary of a document for internal use can be competently generic without anyone losing anything.

Part of the skill is knowing which category you're in. Voice-critical writing — things your audience will actually read and remember, things attached to your reputation, things meant to build a relationship with a reader — requires the extra step of voice preservation. Functional writing — structural scaffolds, first drafts for editing, administrative documents — can use AI output more directly.

Making this distinction deliberately, rather than applying the same workflow to everything, is itself a sign of creative maturity with the tools.

The Takeaway

Your voice is made of specific knowledge and specific choices. The model has neither. Voice preservation is not about fighting the tool — it is about supplying the information the tool lacks, and then doing the rewrite work that no tool can do for you. The result is faster than writing from scratch and more distinctly yours than accepting the first output.

Lesson 2 Quiz

Four questions · Select the best answer for each
1. Why do AI language models tend to produce writing that sounds generic?
Correct. The lesson explains that because models average across millions of texts, they tend toward the median — phrasing that appears in the widest variety of contexts — which is "competent, inoffensive, and utterly unmemorable."
The lesson's explanation is about statistical averaging. The model trends toward the median of its training data — the phrasing that appears most broadly across contexts — rather than toward anything specific or distinctive.
2. What does "specificity injection" mean as a voice-preservation technique?
Correct. Specificity injection means providing your own concrete details — the kind the model cannot invent — so that the output has the specific texture that makes writing feel particular and real.
The lesson's definition of specificity injection is about supplying the concrete details yourself: your mother's kitchen, a specific date, a particular smell. These are things the model cannot invent, but can write around effectively once you provide them.
3. According to the lesson, which type of writing task does NOT require voice preservation effort?
Correct. The lesson categorises meeting agendas as "functional writing" — structural documents that can use AI output more directly because no one is reading them for your distinctive voice or to form a relationship with you.
The lesson distinguishes voice-critical writing (personal essays, reputation-attached content, relationship-building pieces) from functional writing (agendas, summaries, administrative documents). Meeting agendas fall into the latter category.
4. What was Gary Shteyngart's observation about ChatGPT's effect on his writing voice?
Correct. Shteyngart noted that the model made his voice "warmer and less strange" — normalising exactly the idiosyncratic qualities that distinguish his work — leading him to use AI for structural drafts and then rewrite the surface text himself.
Shteyngart's observation was the opposite of experimental difficulty — the model made him more conventional, "warmer and less strange," erasing the strangeness that is central to his literary identity.

Lab 2 — Voice vs. the Median

Diagnose AI blandness and apply voice-preservation techniques · Minimum 3 exchanges

Your Task

In this lab you will practice identifying and correcting AI voice problems. The assistant will generate a deliberately generic AI-style paragraph on a topic you choose, then help you apply one or more voice-preservation techniques to transform it into something that sounds specific and human.

You can also bring your own AI-generated text that you think sounds too generic, and the assistant will help you diagnose which techniques would help most.

Try: "Generate a generic AI-sounding paragraph about [topic], then help me fix it using voice samples and negative constraints." Or: "Here is a paragraph I got from an AI — help me diagnose what's wrong with it and rewrite it with more specific voice."
AI Voice Coach
Lab 2
Welcome to Lab 2. Let's work on the blandness problem. Give me a topic — or paste an AI paragraph you want to improve — and we'll systematically apply the voice-preservation techniques from Lesson 2. I can generate a generic version first so we have something concrete to fix, or we can start with your own material. What's the topic?
Make It Yours: Create With AI · Lesson 3

Making Images and Sound: The Prompt Is Still the Work

How AI image and music generation actually works — and the prompting strategies that make the difference between noise and something useful.
When the output is a picture or a song instead of words, does prompting change — or does the same logic apply?

In August 2022, an image generated by Jason Allen using Midjourney won first place in the digital art category at the Colorado State Fair. The image, "Théâtre D'Opéra Spatial," showed figures in elaborate period costume before a circular portal of light. It beat human artists. Allen had entered it under the name "Jason Allen via Midjourney." The judges did not know it was AI-generated. When they found out, the debate that followed was not primarily about Allen's ethics — it was about what "making" an image means when the making is linguistic: you describe, the model renders.

The uncomfortable fact that the controversy surfaced was this: Allen had spent hours refining his prompts. He was not typing one sentence and accepting the result. He was iterating across dozens of generations, selecting, recombining, adjusting aspect ratios and style references and lighting descriptors. The creative labour was real — it was just invisible, and it was expressed in language rather than in brush strokes or code.

How Image Generation Works

Models like Midjourney, DALL·E 3, and Stable Diffusion are trained on enormous datasets of image-text pairs: a photograph of a red barn at dusk paired with the caption "red barn at dusk, golden hour, pastoral landscape." Through billions of such pairings, the model learns which visual features correspond to which descriptive concepts.

At generation time, the model starts with random noise and iteratively denoises it — removing randomness in the direction that makes the image more consistent with the prompt. The number of denoising steps, the degree of prompt adherence (called "guidance scale" or "cfg"), and the random starting seed all affect the output.

Practically this means: your prompt is a set of instructions to a denoising process. Concepts that appear early in the prompt and are weighted heavily (in Midjourney, by repetition or the :: weighting syntax) have more influence on the output. Vague abstract concepts ("beautiful," "interesting") have less effect than concrete visual descriptions ("diffuse afternoon light coming from camera-left, long shadows").

Image Prompting That Works

Subject
The primary visual element: who or what is in the image. Be specific about species, age, gender presentation, clothing, expression, pose.
Environment
Where and when: interior or exterior, geographic setting, time of day, weather, season. "Overcast winter morning on a Helsinki street" is far more useful than "cold outside."
Lighting
The single most powerful image descriptor. "Harsh overhead fluorescent," "golden hour backlight," "single candle in darkness" each create completely different images from the same subject.
Style reference
A named artist, movement, medium, or era: "in the style of Edward Hopper," "35mm film photograph," "ink wash on rice paper," "brutalist architectural photography." Style references work because they bundle many visual choices into a known package.
Composition and camera
Framing: "extreme close-up on hands," "wide establishing shot," "Dutch angle," "aerial view." Lens language: "shot on 85mm f/1.8," "fisheye distortion." These are borrowed from photography and cinema and they work.
Negative prompts
In Stable Diffusion and some Midjourney versions, you can specify what NOT to include: "no text, no watermarks, no extra limbs, no blurring." This is the image equivalent of constraint prompting in writing.
Documented Result

Adobe's internal research shared at their MAX conference in October 2023 showed that Firefly users who included lighting descriptors and style references in their prompts rated their outputs significantly more satisfactory than those who used subject-only prompts — and required fewer regenerations. Lighting and style reference are the highest-leverage additions to an image prompt.

Music Generation: The Same Logic, Different Grammar

Tools like Suno (released in public beta December 2023) and Udio (launched April 2024) generate music from text prompts. The prompting grammar is different from image generation but the underlying principle is identical: specific descriptors outperform vague ones, and you are guiding a probabilistic system toward a region of its training space.

Effective music prompts typically include: genre ("neo-soul," "Appalachian folk," "Berlin-style techno"), instrumentation ("fingerpicked acoustic guitar, upright bass, no drums"), tempo and feel ("slow, rubato, contemplative"), mood or reference ("melancholic, reminiscent of late-period Talk Talk"), and if using lyric generation, subject matter and emotional arc.

The key limit to understand with current music AI: it is very good at genre and texture, less reliable at structure (verse-chorus-verse following a specific emotional arc), and essentially cannot yet be given a melodic line to harmonise in a controlled way. Understanding this limit tells you where to use it — texture and mood exploration — and where not to — precise compositional structure.

The Through Line

Whether you are prompting for text, images, or audio, the underlying discipline is the same: translate your intention into specific, concrete, technically-informed language that narrows the model's probability space toward the region you actually want. The vocabulary differs by medium. The logic is identical.

Lesson 3 Quiz

Four questions · Select the best answer for each
1. In image generation, which type of descriptor does the lesson identify as the single most powerful?
Correct. The lesson explicitly calls lighting "the single most powerful image descriptor" — noting that different lighting conditions create completely different images even from the same subject.
The lesson specifically identifies lighting as "the single most powerful image descriptor," noting that "harsh overhead fluorescent," "golden hour backlight," and "single candle in darkness" each create completely different results from the same subject.
2. Why do style references (e.g., "in the style of Edward Hopper") work so effectively in image prompts?
Correct. The lesson explains that style references "work because they bundle many visual choices into a known package" — colour palette, brushwork, compositional tendencies, light quality — that the model absorbed during training.
The lesson's explanation is practical: style references work because they package many visual characteristics — colour, light, composition, texture — into a single shorthand that the model learned from its training data.
3. Jason Allen's Midjourney image that won the Colorado State Fair in 2022 was notable because:
Correct. The lesson notes that Allen "spent hours refining his prompts" across dozens of generations — the creative labour was real, just expressed in language rather than traditional mark-making, and invisible to the judges.
The lesson specifically counters the assumption of effortlessness: Allen spent hours iterating, selecting, adjusting. The judges also did not know it was AI-generated — they found out afterward, which triggered the controversy.
4. What is current AI music generation most reliable at, and what is it least reliable at?
Correct. The lesson states that music AI is "very good at genre and texture, less reliable at structure" and "essentially cannot yet be given a melodic line to harmonise in a controlled way" — pointing toward texture and mood exploration as the best current use case.
The lesson's assessment is specific: current music AI excels at genre and texture (the sonic feel of a genre) but is unreliable at following a precise compositional structure or harmonising a given melodic line.

Lab 3 — Image Prompt Engineering

Build and refine image prompts using the six visual components · Minimum 3 exchanges

Your Task

In this lab you will build an image prompt for a visual idea you have, using the six visual components from Lesson 3: Subject, Environment, Lighting, Style Reference, Composition/Camera, and Negative Prompts. The assistant will evaluate your prompt, identify which components are missing or weak, and help you build a stronger version — and then show you what the refined prompt would likely produce.

You can also bring an image prompt that didn't work and we'll diagnose it together.

Try: "I want to create an image of [describe your concept]. Here is my first prompt attempt: [paste your draft]." Or: "Help me build an image prompt from scratch for [your concept] — let's go through the six components one at a time."
Image Prompt Coach
Lab 3
Welcome to Lab 3. Tell me about a visual image you want to create — a concept, a mood, a scene — and share whatever prompt you have so far (even if it's just a rough idea). We'll work through the six visual components to build something that would actually produce what you're imagining. What are we making?
Make It Yours: Create With AI · Lesson 4

What Remains Yours: Authorship, Credit, and Creative Ownership

The practical and ethical questions that arise when AI helped — and how to think about them clearly without the hype in either direction.
If a machine generated the words, the image, or the melody — what did you actually make?

In February 2023, the US Copyright Office issued a registration for Kris Kashtanova's graphic novel "Zarya of the Dawn" — and then immediately walked it back. The Office reexamined the work after public attention revealed that its images had been generated by Midjourney. The final ruling: the text and arrangement of the work could be copyrighted (Kashtanova wrote it and made structural choices), but the individual AI-generated images could not, because copyright requires human authorship of the specific expressive element. In March 2023, the Copyright Office issued further guidance: AI-generated material produced "without creative control" is not copyrightable. Material where a human made "sufficient creative choices" in selection, arrangement, and modification may be.

The legal landscape is still forming — multiple court cases are active as of 2024 — but this ruling established the operational distinction that matters for creators: the process of prompting and selecting is where your authorship lives, and the question is whether that process is sufficient to constitute the "creative control" the law requires.

Three Questions Worth Asking Yourself

The legal and ethical questions around AI authorship do not have clean universal answers yet. But there are three practical questions that help clarify your position in any specific case:

1. What choices did you make?
List them: the concept, the framing, the iteration decisions, the selection from among outputs, the editing, the combination with other work. The more specific and numerous your choices, the stronger your claim to creative authorship of the result.
2. What would not exist without your specific judgment?
If someone else with the same tool could have produced an identical or equivalent result, your distinctive contribution is limited. If your specific knowledge, taste, and iterative decisions were necessary to produce this particular output — that is your contribution.
3. Are you representing it accurately?
Disclosure norms are evolving rapidly by field: journalism has strict requirements, academic work is being codified, commercial creative work has fewer rules but real reputational stakes. Knowing your context and being accurate about what AI contributed is both honest and professionally safer than opacity.

The Attribution Landscape in 2024

Different fields have moved at different speeds. The Associated Press issued AI guidelines in 2023 requiring disclosure whenever AI generated any part of a published work. The National Eating Disorders Association made news for replacing human staff with an AI chatbot in 2023, then reversed course after public criticism — a case study in how disclosure failures compound reputational damage. The Academy of Motion Picture Arts and Sciences ruled in late 2023 that AI-assisted films are eligible for Oscar consideration, but that human creativity must be "predominant."

The working principle that emerges across fields: disclose more rather than less. The cost of over-disclosure is low. The cost of discovered concealment — particularly as detection tools improve — is high. This is not primarily a legal calculation; it is a trust calculation. Your creative reputation rests on what people believe about how you work.

The Training Data Question

An ongoing and unresolved issue: the models were trained on copyrighted material without license in most cases. Getty Images sued Stability AI in January 2023; the New York Times sued OpenAI in December 2023; multiple class-action suits from authors are active. Using these tools does not make you party to those lawsuits, but it does mean the tools exist in contested legal territory. This is worth knowing as a factual matter, separate from whether you choose to use them.

What Collaboration Means

The most useful frame for thinking about authorship when AI is involved may be the one already used for other collaborative situations. When a director makes a film, they did not write the score, operate the camera, or design the costumes — but no one questions whether the film is "theirs." The director's contribution is conceptual, selective, and coordinative. AI-assisted creative work can sit in a similar relationship: the human provides the concept, the taste, the selective judgment, the editorial eye, and the final decision about what is good enough to share.

What this means practically: the more you treat AI as a tool in service of a vision that is clearly yours — where your specific choices, your knowledge of your audience, your aesthetic judgment are the determining factors in the output — the stronger your claim to the creative work, legally, ethically, and in terms of what you can honestly say about how you made it.

The blank page was always a collaboration — with language, with tradition, with influence, with the tools of the moment. AI is a new kind of collaborator, faster and stranger than any previous one. The questions it raises about authorship are not new in kind, only in urgency. You already know how to make creative work that is yours. The discipline is applying that same judgment to a faster, weirder process.

Module Closing Thought

The blank page just got a partner. That partner is fast, tireless, and has no taste. What it produces becomes yours when your specific choices, your knowledge, and your judgment are the deciding factors in what the world finally sees. That responsibility has not changed. Only the speed of the process around it has.

Lesson 4 Quiz

Four questions · Select the best answer for each
1. What did the US Copyright Office rule regarding Kris Kashtanova's AI-illustrated graphic novel "Zarya of the Dawn"?
Correct. The Copyright Office ruled that Kashtanova's text and structural choices could be protected, but the Midjourney-generated images themselves could not — because copyright requires human authorship of the specific expressive element.
The ruling was more nuanced. The Copyright Office split the work: human-authored elements (text, arrangement) were protectable; the AI-generated images themselves were not, because copyright requires human authorship of the specific expressive content.
2. According to the lesson, what is the practical principle that emerges from how different fields are handling AI disclosure?
Correct. The lesson states this explicitly: "The cost of over-disclosure is low. The cost of discovered concealment — particularly as detection tools improve — is high." It frames this as a trust calculation, not just a legal one.
The lesson's working principle is direct: disclose more rather than less. It is not a legal calculation but a trust calculation — your creative reputation rests on what people believe about how you work, and discovered concealment damages that.
3. The lesson compares AI-assisted creative authorship to which existing creative role?
Correct. The lesson uses the film director analogy: the director didn't write the score, operate the camera, or design the costumes — but the film is unquestionably theirs because their conceptual, selective, and coordinative choices determined the outcome.
The lesson's analogy is the film director — someone whose authorship is real and unquestioned even though they did not personally execute every craft element. The director's vision, selection, and judgment are what make the film theirs.
4. The US Copyright Office's March 2023 guidance established that AI-generated material may be protectable when:
Correct. The Copyright Office's language was specifically about "sufficient creative choices" — selection, arrangement, and modification — establishing that the locus of protectable human authorship is in those decisions, not merely in the act of prompting.
The Copyright Office's guidance was specifically about human creative control: material where a human made "sufficient creative choices" in selection, arrangement, and modification may be protected. The tool's origin or the user's payment terms are not relevant to this test.

Lab 4 — Authorship Audit

Map your creative choices in an AI-assisted project · Minimum 3 exchanges

Your Task

In this lab you will work through an authorship audit of a real or hypothetical AI-assisted creative project. The assistant will ask you structured questions about your creative choices — what you decided, what you selected, what you changed — and help you build a clear account of where your authorship sits and how you would disclose AI's role accurately in your specific context.

This is also where you can bring harder questions: what to say in an artist's statement, how to handle a client who asks about AI, what disclosure looks like in your field.

Try: "I used AI to help create [describe the project]. Walk me through an authorship audit so I understand what is mine and how to disclose it." Or: "I'm in [field] and I need to know how to handle AI disclosure — what are the current norms?"
Authorship Advisor
Lab 4
Welcome to Lab 4. Let's do an authorship audit. Tell me about a creative project you made (or are planning to make) with AI assistance — what it is, what the AI generated, and what choices you made. I'll help you map your creative contributions and work out how to represent the project's authorship accurately in your specific context. What's the project?

Module Test

15 questions across all four lessons · 80% to pass
1. What is the primary reason AI language models tend toward generic output?
Correct. The model produces the most probable continuation of the context — which means it trends toward the median of its training data.
The lesson attributes generic output to statistical averaging across training data — the model produces what is most probable across the widest variety of contexts.
2. Which prompt component involves giving the model one or two examples of the output style you want?
Correct. Examples — also called few-shot prompting — involves including sample outputs in the prompt so the model can align to your target style.
This is the Examples component, also called few-shot prompting — and the lesson calls it "often the single highest-value addition to a prompt."
3. Kazuo Ishiguro's argument about AI and novel-writing centred on:
Correct. Ishiguro argued that novel-writing is an act of discovery — the author confronts things they don't yet know — and that AI, producing probable continuations, cannot do this.
Ishiguro's point was philosophical: a novel requires genuine not-knowing, and AI produces statistically expected output rather than genuine discovery.
4. The iterative prompt engineering method is best described as:
Correct. The lesson describes the discipline as run, evaluate, diagnose, refine — treating each output as material to react to rather than a final result.
The lesson specifically describes iteration as: run, evaluate, diagnose, refine. The first output is a draft, not a conclusion.
5. According to the lesson on voice, what is the recommended treatment of AI output for voice-critical writing?
Correct. The rewrite layer is the recommended approach: accept the structure and key ideas from the AI, then rewrite the surface text in your own words.
The lesson recommends using AI output as structural scaffold, then rewriting the surface text yourself — fast because the structural work is done, and distinctly yours because the sentences are.
6. "Negative constraints" in a writing prompt means:
Correct. Negative constraints ban specific elements: "Do not use the word innovative, do not make ROI claims, keep sentences under 20 words."
Negative constraints are about exclusion — explicitly listing what the model should NOT include or do — and the lesson notes this is especially effective at reducing AI fingerprinting phrases.
7. In image prompting, what does the "guidance scale" or "cfg" parameter control?
Correct. The guidance scale determines how closely the denoising process follows the prompt — higher values mean more prompt adherence, lower values allow more model creativity.
The lesson defines guidance scale as the "degree of prompt adherence" — how closely the model's denoising process follows your specific prompt instructions.
8. Jason Allen's Colorado State Fair win with a Midjourney image was controversial primarily because:
Correct. The judges discovered after the award that the image was AI-generated — triggering a debate about what constitutes making an image when the work is expressed in language rather than mark-making.
The controversy was specifically about the undisclosed AI generation — the judges only discovered it afterward — and the deeper question it raised about whether linguistic creative labour (prompting) counts as making an image.
9. What is current AI music generation LEAST reliable at?
Correct. The lesson identifies precise compositional structure and melodic harmonisation as current weak points, steering users toward texture and mood exploration as the stronger use case.
The lesson specifically identifies structural precision and melodic harmonisation as current limitations — the tools excel at genre texture and mood but cannot reliably follow a prescribed compositional arc.
10. The US Copyright Office's key test for whether AI-assisted work can receive copyright protection is:
Correct. The Copyright Office's March 2023 guidance specifically turned on "sufficient creative choices" in selection, arrangement, and modification — making those acts of human judgment the locus of protectable authorship.
The Copyright Office's test is about human creative control: "sufficient creative choices" in selection, arrangement, and modification. No percentage thresholds or licensing conditions are part of the legal test.
11. The lesson compares the "Role" component of a prompt to:
Correct. Role defines who the model should act as, which shifts its default register, vocabulary, and assumptions — for example, "You are a copywriter who specialises in B2B software."
Role is specifically about persona — who the model should act as — which in turn shifts its vocabulary, register, and default assumptions about the task.
12. Gary Shteyngart's experience with ChatGPT and his writing voice showed that:
Correct. Shteyngart observed that the model made his voice "warmer and less strange" — erasing the idiosyncratic qualities central to his literary identity — leading him to reserve AI for structural work and rewrite the surface text himself.
Shteyngart's finding was the opposite of accurate replication — the model normalised his voice, making it more conventional and less distinctively his.
13. Which organisation sued OpenAI in December 2023 over training data?
Correct. The New York Times filed suit against OpenAI in December 2023 over training data. Getty Images sued Stability AI in January 2023.
The lesson mentions multiple suits: Getty Images sued Stability AI in January 2023; the New York Times sued OpenAI in December 2023. These are distinct cases involving different organisations and companies.
14. The lesson's authorship audit asks you to consider: "What would not exist without your specific judgment?" This question is designed to identify:
Correct. The question tests whether your specific contribution was necessary — if anyone with the same tool could have produced an equivalent result, your distinctive contribution is limited.
The audit question probes whether your specific choices, taste, and knowledge were necessary to produce this particular output — or whether any user of the same tool would have arrived at something equivalent.
15. What does the Academy of Motion Picture Arts and Sciences' late 2023 ruling on AI and Oscar eligibility establish?
Correct. The Academy ruled that AI-assisted films remain eligible for Oscar consideration, provided that human creativity is "predominant" — establishing the same general principle as the Copyright Office: human creative primacy matters.
The Academy's ruling was that AI-assisted films remain eligible, but human creativity must be "predominant." This aligns with the broader pattern across fields: AI involvement is permitted, but human creative primacy is the threshold condition.