The Credits Are Getting Longer: Who Gets Named When a Machine Helped Make It?
Films have end credits, albums have liner notes, books have acknowledgments. Generative tools are pushing every creative field to rethink what 'made by' means, and some artists are writing new rules.
By Priya Sethi
Creative AI Editor
The closing credits of The Lamplighter's Daughter, an eleven-minute animated short by Moth & Marrow, a seven-person co-op in Leeds, run for exactly ninety seconds. Most of it is familiar: direction, animation, voices, music, and a thank-you to someone's mum for the loan of a garage.
Then, near the end, comes one line that made a festival audience murmur: Background textures: generated, then repainted by hand by Ade Bramwell.
"We argued about that line for three weeks," says Cass Mbeki-Lowe, who directed the film. "Not about whether to include it. About the verb."
Generated, then repainted. It's a short sentence that carries a big question, and every creative field is now trying to answer it in its own way. When a machine helped make something, what does "made by" mean, and who gets named?
Credits were always a social contract
It's easy to forget that credits were invented, and fairly recently in some fields. Early films often had no end credits at all. Album liner notes grew from a marketing afterthought into a cherished record of who played what. Book acknowledgments evolved from dedications to wealthy patrons into long, grateful lists of the editors, agents and partners who put up with the writer.
Each of these formats exists because audiences and collaborators wanted the same thing: an honest account of how the work got made, and fair recognition for the people who made it. They're mostly not legal documents. They're a promise about the truth.
Generative tools didn't create the question of what counts as authorship. Photographers argued about darkroom manipulation, musicians argued about sampling and illustrators argued about tracing. What's new is the urgency, because a generative contribution can be large, invisible and hard to describe.
A credit isn't a confession or a legal shield. It's the oldest promise in creative work: this is how it was made, and these are the people who made it.
Three ways artists are labelling
Over the past two months I've talked to a dozen artists, designers and musicians, and three main approaches are emerging. None of them is standard yet.
The ingredient label. This is a list of tools and what each one did, like the back of a cereal box. Ottoline Varga, the illustrator and composer, prints one in her album booklets. It's clear and easy to scan, and it treats generative tools the same way as a piano or a dip pen, as instruments.
The process note. This is a short paragraph explaining the workflow in plain language. Moth & Marrow's website has one for every film, covering what was storyboarded by hand, where generative tools were tried and rejected, and which textures survived and how they were repainted. It's more work, but it tells a story that audiences often find fascinating.
The percentage. Some creators have tried putting a number on it, such as "about 15% AI-assisted." It's tempting because it looks precise, but almost everyone I spoke to who tried it gave it up. Fifteen percent of what? Time? Pixels? Decisions? A single generated texture can define a film's mood, while a thousand generated drafts can contribute nothing to the final cut. Percentages measure the wrong thing very exactly.
Disclosure as a creative choice
It's commonly assumed that labelling AI involvement is purely defensive, something you do to avoid backlash. The artists I spoke to describe something more interesting.
Kasimir Lund, a freelance editorial illustrator, lost a regular client last year after it emerged that he'd used a generative tool for background patterns without saying so. "They weren't angry about the tool," he says. "They were angry that they found out from someone else."
He now includes a two-line process note with every delivery, and he says his commissions have increased. Part of that is because art directors appreciate knowing exactly what they're buying. Part is because the notes show his judgement. "The note says: I tried this, I threw that away, I drew this part myself because it mattered. That's the job, really. Choosing."
That shift is worth noticing. A good credit doesn't just disclose; it reveals taste. It shows where the artist decided their own hand was non-negotiable.
The credits a model can't give
There's a harder part of this conversation, and it would be dishonest to skip it. Generative models learn from enormous amounts of existing work, much of it made by artists who were never asked and are never named. How to handle that, including consent, compensation and the ability to opt out, is being argued about in studios, courts and legislatures around the world, and it's far from settled.
A credits line can't solve that. But several of the artists I spoke to said it shapes which tools they choose. Some prefer tools trained on licensed or openly contributed material, where the provider is clear about its sources. Some train small models only on their own back catalogue. Some avoid image generators entirely while happily using AI for transcription or colour correction, where the question feels different.
"I can't credit the thousands of painters whose work taught a model what a sky looks like," says Cass. "But I can choose my tools like their opinions matter."
A template you can steal
If you're making something with generative help and want to label it, here's a simple structure that borrows the best of the ingredient label and the process note:
- Made by: the humans, with their roles.
- Made with: the tools, generative and otherwise.
- What the machine did: in plain verbs, such as drafted, generated, transcribed, upscaled or suggested.
- What people did: in equally plain verbs, such as chose, repainted, rewrote, rejected or performed.
- What was never automated: the part you'd defend in an argument.
Audiences read that last line most closely, and it says the most about you.
At Ideabro we label every article in a similar spirit. You'll see a small note on each piece saying whether it was written by a person, assisted by AI, or generated and then edited. It's an experiment, and we'll keep refining it. But it started from the same instinct as Moth & Marrow's three-week argument over a single verb: readers deserve to know how the thing in front of them was made.
Ade Bramwell, the painter who repainted those generated textures, gets the last word. "Put my name next to the verb that took the longest," he says. "That's all anyone wants from a credit."
How this was made Human-written
Human-written. Reported and written by a person. AI not used for the text.
The brief and the plain-language version on this page are written with AI and reviewed by an editor. The story itself is the source of truth.
Ask about this story
Conversation0
No comments yet.
Start the conversation — the first reply sets the tone.