The Supreme Court ruled this week on the question every AI artist has been asking: Can AI-generated art be copyrighted?

The answer: Yes, but only if there’s “significant human creative input.”

Which raises the obvious question: What counts as “significant”?

The Case That Got Here

The lawsuit involved an artist who used Midjourney to generate a series of images, then edited them extensively in Photoshop. She registered the final works with the Copyright Office. They said no. She sued. She won. The publisher appealed. And now the Supreme Court has spoken.

The Ruling (In Plain English)

Justice Kavanaugh’s majority opinion (6-3) walked a tightrope:

“The Copyright Act protects ‘original works of authorship.’ Authorship requires a human author. But human authorship can incorporate AI-generated elements, provided the human’s creative contribution is significant and the AI’s role is instrumental rather than determinative.”

Translation: If you prompt “cyberpunk cityscape” and Midjourney does all the work, you’re not the author. If you generate 500 variations, composite 12 of them, paint over them, add original elements, and create something new — you might be.

The Problem Nobody’s Talking About

Nobody knows what “significant” means.

  • Is selecting 1 out of 100 AI outputs enough?
  • Does color correction count?
  • What about arranging AI elements into a collage?
  • How about training a LoRA on your own art, then generating with it?
  • Is inpainting “significant”? What about outpainting?
  • Does the number of iterations matter?

Lawyers interviewed for this piece used phrases like “fact-specific inquiry” and “case-by-case analysis.” Which is legal speak for “we have no idea, see you in court.”

What This Means for Different Groups

For AI Artists

Document everything.

Save your layers. Keep your prompts. Screenshot your generation process. Note which elements you modified and how. You may need to prove your human contribution in court someday.

Best practices:

  • Save all intermediate generations
  • Keep Photoshop layer files
  • Document your prompt iterations
  • Note which elements are AI vs. human-created
  • Consider registering works with detailed deposition statements

For AI Companies

This is a win.

The Court didn’t ban AI art. It didn’t say AI companies are liable for user creations. It just said humans need to be meaningfully involved in the final work.

Midjourney, Stable Diffusion, and DALL-E can keep operating. But expect them to add more documentation features to help users prove “significant human input.”

Cha-ching.

Expect a decade of litigation to define “significant.” Every edge case will end up in court:

  • AI-assisted photography
  • AI music production
  • AI-written code
  • AI-generated 3D models
  • AI-edited video

Each will require its own “fact-specific inquiry.”

The Precedents That Matter

The Court cited three key cases:

  1. Burrow-Giles Lithographic Co. v. Sarony (1884) — Photographs can be copyrighted because photographers make creative choices (lighting, pose, composition)

  2. Feist Publications v. Rural Telephone Service (1991) — Pure labor (“sweat of the brow”) isn’t enough; there must be creative authorship

  3. Zarya of the Dawn (2023) — Copyright Office’s recent decision denying registration for AI-generated comic panels

The throughline: Human creativity matters more than human labor.

The International Angle

This ruling applies to U.S. copyright law. Other jurisdictions are still figuring it out:

  • EU: The AI Act is silent on copyright, focusing on transparency
  • UK: Currently allows computer-generated works, but reviewing the policy
  • China: Recently ruled AI images CAN be copyrighted (Beijing Internet Court, 2023)
  • Japan: Generally permissive of AI training and generation

Expect forum shopping. An AI artist denied copyright in the U.S. might seek protection in China or the UK.

What Comes Next

Short term (2026):

  • First wave of lawsuits testing the boundaries
  • Copyright Office issues guidance (probably vague)
  • AI tools add documentation features

Medium term (2027-2028):

  • Circuit courts split on what “significant” means
  • More Supreme Court cases likely
  • Industry standards emerge (maybe)

Long term (2030+):

  • Congressional action possible (new copyright amendments)
  • International treaties may address AI authorship
  • Or we just… figure it out through case law

The Real Story

This ruling doesn’t settle the debate. It just moves it from “can AI art be copyrighted?” to “what counts as human enough?”

And that’s a much messier question.

Because “significant human input” isn’t a technical standard. It’s a philosophical one. It’s asking: What makes art human?

Is it the choices? The labor? The intention? The mistakes?

The Court didn’t answer that. It just kicked the can down the road.

And now it’s up to artists, lawyers, and judges to figure it out — one lawsuit at a time.


The plot twist? We’re about to learn more about human creativity from AI copyright cases than from a century of art theory.

Because sometimes you have to define what something ISN’T before you understand what it IS.


This article is part of Plot Twist Daily’s AI coverage. Follow @PlotTwist_Daily for more.

Related: