Writers and artists need a way to label their use of AI: What could this look like?

Today’s post is by poet and filmmaker Dave Malone, with AI assistance (editing and drafting assistance) by Claude AI and editing by PGB (human editor).

---

I was curious to see if Claude AI could write poems as good as mine. We conducted an experiment on TikTok in September. I shared two poems titled “Autumn Witness.” One I wrote, and the other was written by Claude AI after I shared my work and asked him to write in my voice and style. I asked my followers to guess which poem was mine.

76% were wrong. They chose a poem by Claude.

The comments revealed something that may be disturbing to some people. “I resonated more with #1, more positive, more spiritual,” one follower wrote of Claude’s poem.

Another said with complete confidence: Some of my followers declare the opposite assurance. “#1 is Claude using a common phrase. What makes the helicopter leaf and altar match in #2 is the human spirit connecting disparate similarities.” Another person revealed a hard truth: “It’s frustrating that it’s this difficult.”

A week later, I conducted a similar experiment at the Ozark Symposium in West Plains, Missouri. 59% thought Claude wrote my poem.

These experiments were born out of simple curiosity, but I started thinking about more complex problems. If AI can create art as captivating as humans, audiences certainly deserve transparency. Similarly, artists need to be transparent about their use of AI.

I’m a poet and playwright from the Ozarks, Missouri, and I’ve been interested in AI from the beginning. For me, Claude started out as a research assistant, but I called him a “Google machine on steroids.” Things have changed. I use Claude almost every day in my writing practice, and he has become a valuable sounding board as well as an editor for my poems, screenplays, film projects, book submissions, and this article you’re reading right now. I went from being an AI enthusiast to an AI active person, and I know that transparency is important.

The creative industries lack agreed standards for transparency around the use of AI. Artists don’t know how to publish their usage, and then the audience doesn’t know whether they’re using AI or human work. Some creators are upfront about their use of AI, while others keep it a secret. Many are somewhere in the vague middle.

At the moment, there are no clear standards, so artists don’t know what’s expected of them, and audiences don’t know what they’re going to get. Considering this problem, we created a solution.

Two categories of solutions

AACC (AI Attribution and Creative Content) is an open source transparency framework with two simple categories.

AI support This means that creators invent and promote their work, and that AI contributes to the process.

AI-generated This means that the AI ​​creates the primary content based on the creator’s prompts and instructions, and the creator’s role is conceptual and editorial.

that’s it. Two categories broadly cover how artists work with AI.

AI support

• Creators create and promote works
• AI contributes through generation, modification, and enhancement at every step.
• Creators make all major creative decisions and are responsible for the final product

Authors are already working this way, even without a formal label. In a recent BookBub survey of authors (69% self-published, 6% traditional, 25% both self-published and traditional), some authors described using AI as an editor for “research, grammar editing, and sometimes paraphrasing,” or as a developmental editor to “talk about plot, play around with character profiles, and work with structural templates.”

One person described using AI to “draft sentences, scene descriptions, and rewrite sentences.” The author emphasized that their use was not a “blind” generation, but material that was “considered/accepted/rejected to remain within the scope of my opinion and vision.”

AI-generated

• AI creates primary content based on creator prompts and instructions
• The creator’s role is conceptual, curation, and editing.

Another author surveyed explained, “Sometimes I think of myself as an assistant or even a ghostwriter…I come up with an idea and play with the AI ​​to make it better. Then the AI ​​purges the first draft and I take over as author from there.”

These two AACC categories run the gamut from using AI to refine your writing to generating an entire first draft. Lack of this kind of consistent labeling creates problems.

Why you need this framework: Examples from mainstream artists

AI-assisted novel: Dojojima, Tokyo by Rie Kudan

In 2024, author Rie Kudan admitted (after the fact) that 5% of his novel was “taken straight from ChatGPT.” Would the conversation about her award-winning work have changed if she had included a transparent label from the beginning?

AI-assisted movies: Irish

For this 2019 Martin Scorsese film, editors used AI deepfake software FaceSwap to make the film’s leads, Robert De Niro, Al Pacino, and Joe Pesci, look younger. While movie audiences are largely receptive to AI-driven visual effects, I argue that transparent disclosure should still be the norm.

AI-assisted or AI-generated music: Telisha “Nicky” Jones

Recently, Telisha “Nicky” Jones, who plays Zania Mone, signed a $3 million record deal with music created using AI. She writes the lyrics, but Suno’s AI handles the vocals. Is her music AI-driven or AI-generated?

AI-generated music: Velvet Sundown

The Velvet Sundown initially streamed their music on Spotify without knowing how they made their music. They have since updated their Spotify profile to say that their music project was “composed, audio-visualized, and visualized with the support of artificial intelligence.” These keywords refer to works generated by AI.

AI-generated novel: death of the author Written by Stephen Marche

In an interview with Joanna Penn, Stephen Marche made clear the use of AI, which makes up 95% of the novel. Death of the author. Marche described his role as a “curator” and himself as a “creator” of the work, which would be classified as AI-generated due to the extensive use of AI.

Without a consistent framework, each artist will respond differently when it comes to making AI disclosures. Audiences often have no idea what role AI played in the production. Artists don’t know what is expected of them. That’s the problem this framework solves.

move forward

To be sure, the AACC framework doesn’t solve all the problems with AI in creative work. Like many artists, I have ethical concerns about how most AI models are trained. Major companies (with exceptions such as Apertus) were using copyrighted works without the consent or compensation of the artists. Data centers have environmental costs. These are real issues.

But the hard truth is that, just as illegal file sharing through Napster eventually evolved into licensed platforms like Spotify and Pandora, the current wave of lawsuits against AI companies will likely result in proper licensing and payment systems being established. It’s a long game. In the meantime, artists will have to make an immediate choice: use AI transparently or leave their audience guessing.

For young writers and artists, I echo musician Nick Cave’s warning against leaning too much towards AI generation in the arts. Don’t avoid the “inconvenience of artistic struggle” by going straight to easy products. That struggle is where you develop your voice and your art. There are no shortcuts. AI is a tool, and like any tool, it reveals the skill of the person using it. An inexperienced curator creates an inexperienced work.

For those of us who use AI as a legitimate tool, transparency is not complicated. Label your work. Help your audience understand your process. Use a framework or write your own, but be honest about how you work.

The AACC framework is open source and available on GitHub and my website. Two simple categories with clear attribution. Whether you’re curious about AI, are positive about it, or are still figuring it out, we encourage you to be transparent. It’s that simple.