# AI use opens Hollywood to public domain (?)

**URL:** <https://boards.straightdope.com/t/ai-use-opens-hollywood-to-public-domain/990549>\
**Category:** Great Debates\
**Tags:** movies, ai\
**Created:** [September 22, 2023, 3:56pm UTC](https://boards.straightdope.com/t/ai-use-opens-hollywood-to-public-domain/990549 "2023-09-22T15:56:58Z")\
**Posts on this page:** 20\
**Page:** 1

<div class="post-metadata">

**Author:** ![Sage\_Rat](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/sage_rat/32/399_2.png) [@Sage\_Rat](https://boards.straightdope.com/u/Sage_Rat)\
**Post date:** [September 22, 2023, 3:56pm UTC](https://boards.straightdope.com/t/ai-use-opens-hollywood-to-public-domain/990549/1 "2023-09-22T15:56:58Z")

</div>

The US copyright office and the judicial system has recently decided that [AI generated art cannot be copyrighted](https://www.reuters.com/legal/ai-generated-art-cannot-receive-copyrights-us-court-says-2023-08-21/). Ergo, any AI art is de facto in the public domain.

Right now, Hollywood is having a fight over the use of AI for movie production. But, based on the above decision, any such use would seem to mean that their products would - at least in part - be put into the public domain. Any segments of film, any characters, any plot elements that had been AI generated would end up out of copyright and everyone else would be free to take and use those elements in their own films, put on t-shirts, make toys, etc.

Is this understanding correct?

---

<div class="post-metadata">

**Author:** ![TriPolar](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/tripolar/32/3008_2.png) [@TriPolar](https://boards.straightdope.com/u/TriPolar)\
**Post date:** [September 22, 2023, 4:35pm UTC](https://boards.straightdope.com/t/ai-use-opens-hollywood-to-public-domain/990549/2 "2023-09-22T16:35:57Z")

</div>

How much human input would be needed to make it copyrightable? I don’t think anyone is planning to make a movie based completely on AI output. Character design, plot elements will be tweaked by humans, and they were likely created based on human input to start with.

I think this idea is silly. You can copyright a picture by pointing your phone in a random direction and tapping your screen. How is running an AI app on that same phone any different?

---

<div class="post-metadata">

**Author:** ![Sage\_Rat](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/sage_rat/32/399_2.png) [@Sage\_Rat](https://boards.straightdope.com/u/Sage_Rat)\
**Post date:** [September 22, 2023, 5:29pm UTC](https://boards.straightdope.com/t/ai-use-opens-hollywood-to-public-domain/990549/3 "2023-09-22T17:29:32Z")

</div>

> [@TriPolar](#):
>
> Character design, plot elements will be tweaked by humans, and they were likely created based on human input to start with.

To what extent is “tweaking” necessary to lock the element in? How does the company demonstrate that they have done so? How strong of evidence should an outside party have before deciding that they can proceed with use of those elements?

I’d generally expect that the standard would be that the company has successfully submitted the element for copyright. The copyright office would need to see sufficient evidence of human involvement for all elements.

---

<div class="post-metadata">

**Author:** ![Cervaise](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/cervaise/32/16693_2.png) [@Cervaise](https://boards.straightdope.com/u/Cervaise)\
**Post date:** [September 22, 2023, 5:32pm UTC](https://boards.straightdope.com/t/ai-use-opens-hollywood-to-public-domain/990549/4 "2023-09-22T17:32:57Z")

</div>

Relevant recent court case, for context.

> **[US rejects AI copyright for famous state fair-winning Midjourney art](https://arstechnica.com/information-technology/2023/09/us-rejects-ai-copyright-for-famous-state-fair-winning-midjourney-art/)**
>
> Controversial AI art piece from 2022 lacks human authorship required for registration.

Things are going to get a lot muddier before they become clearer.

---

<div class="post-metadata">

**Author:** ![TriPolar](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/tripolar/32/3008_2.png) [@TriPolar](https://boards.straightdope.com/u/TriPolar)\
**Post date:** [September 22, 2023, 6:36pm UTC](https://boards.straightdope.com/t/ai-use-opens-hollywood-to-public-domain/990549/5 "2023-09-22T18:36:08Z")

</div>

> [@Sage\_Rat](#):
>
> To what extent is “tweaking” necessary to lock the element in? How does the company demonstrate that they have done so? How strong of evidence should an outside party have before deciding that they can proceed with use of those elements?

It should be pretty easy to record the input provided to the machine to produce the starting version of the work. The programs used to create the works would likely have accompanying editors for post-gen changes. A modern copyrighted work might include the entire object instance data from start to finish including planning stage discussions from a project app.

You can copyright images from your Ring doorbell. Why can’t you copyright images that your phone randomly creates?

---

<div class="post-metadata">

**Author:** ![DPRK](https://avatars.discourse-cdn.com/v4/letter/d/4491bb/32.png) [@DPRK](https://boards.straightdope.com/u/DPRK)\
**Post date:** [September 22, 2023, 6:48pm UTC](https://boards.straightdope.com/t/ai-use-opens-hollywood-to-public-domain/990549/6 "2023-09-22T18:48:19Z")

</div>

> [@TriPolar](#):
>
> You can copyright images from your Ring doorbell. Why can’t you copyright images that your phone randomly creates?

There have always been “sweat of the brow”, “modicum of creativity” and other guidelines. So you could not copyright random images from your Ring doorbell.

---

<div class="post-metadata">

**Author:** ![TriPolar](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/tripolar/32/3008_2.png) [@TriPolar](https://boards.straightdope.com/u/TriPolar)\
**Post date:** [September 22, 2023, 7:42pm UTC](https://boards.straightdope.com/t/ai-use-opens-hollywood-to-public-domain/990549/7 "2023-09-22T19:42:40Z")

</div>

So how much sweat and how large a modicum is necessary?

---

<div class="post-metadata">

**Author:** ![wolfpup](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/wolfpup/32/10618_2.png) [@wolfpup](https://boards.straightdope.com/u/wolfpup)\
**Post date:** [September 22, 2023, 8:05pm UTC](https://boards.straightdope.com/t/ai-use-opens-hollywood-to-public-domain/990549/8 "2023-09-22T20:05:35Z")

</div>

Allen says he submitted hundreds of prompts to guide the image creation process, and then completed the editing with Photoshop. Photoshop alone is powerful enough to completely change the content and character of an image. If accurate, this sounds like he put a lot of time and effort into the creative process, and was essentially just using Midjourney as a very sophisticated rendering tool. Obviously I have no idea how closely his account corresponds to the reality, but there’s at least the possibility that the hard line the Copyright Office is taking may be excessively prohibitive against legitimate copyright claims.

---

<div class="post-metadata">

**Author:** ![Sage\_Rat](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/sage_rat/32/399_2.png) [@Sage\_Rat](https://boards.straightdope.com/u/Sage_Rat)\
**Post date:** [September 22, 2023, 9:22pm UTC](https://boards.straightdope.com/t/ai-use-opens-hollywood-to-public-domain/990549/9 "2023-09-22T21:22:37Z")

</div>

I think there’s certainly an argument that a sufficiently advanced description should be copyrightable, especially if it can be shown to consistently produce good and identifiable outputs.

That would just be the text description, though.

I’d be curious whether the copyright office saw the before and after of his “cleanup” work and judged on that or simply said, “AI is no go.”

---

<div class="post-metadata">

**Author:** ![Novelty\_Bobble](https://avatars.discourse-cdn.com/v4/letter/n/a3d4f5/32.png) [@Novelty\_Bobble](https://boards.straightdope.com/u/Novelty_Bobble)\
**Post date:** [September 23, 2023, 4:07pm UTC](https://boards.straightdope.com/t/ai-use-opens-hollywood-to-public-domain/990549/10 "2023-09-23T16:07:52Z")

</div>

Something that perhaps others more knowledgeable can clear up.

If I arrive at a publishers with a completed story/screenplay/script, and I say I created it, how can anyone ever prove that false?

---

<div class="post-metadata">

**Author:** ![echoreply](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/echoreply/32/3641_2.png) [@echoreply](https://boards.straightdope.com/u/echoreply)\
**Post date:** [September 23, 2023, 5:39pm UTC](https://boards.straightdope.com/t/ai-use-opens-hollywood-to-public-domain/990549/11 "2023-09-23T17:39:22Z")

</div>

This is another one of those questions, where the very unsatisfying answer is “whatever the courts decide.”

Copyright only applies to human creations. That is the law as it stands now. Perhaps in the future it will change. How much human input is necessary for copyright to attach?

A human setup the camera, but a monkey pressed the shutter button. No copyright.

A human set out the paper and paints, but an elephant painted the picture. No copyright.

A human provided the prompts, but an AI drew the picture. No copyright.

---

<div class="post-metadata">

**Author:** ![Alessan](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/alessan/32/465_2.png) [@Alessan](https://boards.straightdope.com/u/Alessan)\
**Post date:** [September 23, 2023, 5:55pm UTC](https://boards.straightdope.com/t/ai-use-opens-hollywood-to-public-domain/990549/12 "2023-09-23T17:55:18Z")

</div>

> [@Novelty\_Bobble](#):
>
> If I arrive at a publishers with a completed story/screenplay/script, and I say I created it, how can anyone ever prove that false?

If AI is good enough to write a marketable story/screenplay/script, why would publishers be accepting submissions? They could just pump them out themselves.

---

<div class="post-metadata">

**Author:** ![Voyager](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/voyager/32/133_2.png) [@Voyager](https://boards.straightdope.com/u/Voyager)\
**Post date:** [September 23, 2023, 7:14pm UTC](https://boards.straightdope.com/t/ai-use-opens-hollywood-to-public-domain/990549/13 "2023-09-23T19:14:57Z")

</div>

> [@Novelty\_Bobble](#):
>
> If I arrive at a publishers with a completed story/screenplay/script, and I say I created it, how can anyone ever prove that false?

They might be able to run it through software that looks for the telltale signs of AI generation. And no doubt you soon will be required to sign something saying it was your own work. They can’t prove it wasn’t plagiarized either, but the consequences of them finding out are severe.

---

<div class="post-metadata">

**Author:** ![Darren\_Garrison](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/darren_garrison/32/92_2.png) [@Darren\_Garrison](https://boards.straightdope.com/u/Darren_Garrison)\
**Post date:** [September 23, 2023, 8:13pm UTC](https://boards.straightdope.com/t/ai-use-opens-hollywood-to-public-domain/990549/14 "2023-09-23T20:13:02Z")

</div>

> **[Opinion: The Copyright Office is making a mistake on AI-generated art](https://arstechnica.com/tech-policy/2023/09/opinion-dont-exclude-ai-generated-art-from-copyright/amp/)**
>
> A rule against copyrighting AI art will be unworkable.

---

<div class="post-metadata">

**Author:** ![DrDeth](https://avatars.discourse-cdn.com/v4/letter/d/b487fb/32.png) [@DrDeth](https://boards.straightdope.com/u/DrDeth)\
**Post date:** [September 23, 2023, 8:31pm UTC](https://boards.straightdope.com/t/ai-use-opens-hollywood-to-public-domain/990549/15 "2023-09-23T20:31:44Z")

</div>

> [@Sage\_Rat](#):
>
> Ergo, any AI art is de facto in the public domain.

Nope. It must be _100% AI generated_, and that is “only” a District court ruling, which can be overturned.

> [@TriPolar](#):
>
> How much human input would be needed to make it copyrightable? I

Looks like about 1%.

_A work of art created by artificial intelligence without any human input cannot be copyrighted under U.S. law, a U.S. court in Washington, D.C., has ruled._ Note 'any".

But yeah, this ruling will make AI generated stuff less appealing.

---

<div class="post-metadata">

**Author:** ![TriPolar](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/tripolar/32/3008_2.png) [@TriPolar](https://boards.straightdope.com/u/TriPolar)\
**Post date:** [September 24, 2023, 4:11pm UTC](https://boards.straightdope.com/t/ai-use-opens-hollywood-to-public-domain/990549/16 "2023-09-24T16:11:08Z")

</div>

> [@TriPolar](#):
>
> How much human input would be needed to make it copyrightable? I

This brings up a host of questions.

First, what is AI? To make any sense this ruling should apply to random products created by any machine, and I think that is already established. “AI” does not have a clear and well accepted definition and certainly shouldn’t be used in any court judgement.

Next, even with future products which could be better qualified as AI than the current parlor game technology there has to be a means to measure human input relative to the entire work. Will it be a question of time, percentage of the work, or some other calculation? And the toughest part will be in evaluating how much creative effort was supplied by humans as opposed to asking the program to produce a romcom script which will actually be based on existing romcoms instead of anything new and original. Not really different than the way scripts are produced by humans, but the originality of a large work that doesn’t reach the level of plagiarism is considered copyrightable if a human produced it.

---

<div class="post-metadata">

**Author:** ![DrDeth](https://avatars.discourse-cdn.com/v4/letter/d/b487fb/32.png) [@DrDeth](https://boards.straightdope.com/u/DrDeth)\
**Post date:** [September 24, 2023, 8:00pm UTC](https://boards.straightdope.com/t/ai-use-opens-hollywood-to-public-domain/990549/17 "2023-09-24T20:00:28Z")

</div>

> [@TriPolar](#):
>
> This brings up a host of questions.
> 
> First, what is AI? To make any sense this ruling should apply to random products created by any machine, and I think that is already established. “AI” does not have a clear and well accepted definition and certainly shouldn’t be used in any court judgement.

Yeah, the Wizards of the Coast haters got all het up because one of the artists admitted finishing some of his art with AI… without WotC’s knowledge. I guess their idea was that no one should use any AI art at all.

---

<div class="post-metadata">

**Author:** ![OldGuy](https://avatars.discourse-cdn.com/v4/letter/o/3bc359/32.png) [@OldGuy](https://boards.straightdope.com/u/OldGuy)\
**Post date:** [September 24, 2023, 10:50pm UTC](https://boards.straightdope.com/t/ai-use-opens-hollywood-to-public-domain/990549/18 "2023-09-24T22:50:53Z")

</div>

> [@Voyager](#):
>
> And no doubt you soon will be required to sign something saying it was your own work. They can’t prove it wasn’t plagiarized either, but the consequences of them finding out are severe.

Actually publishers have required this for year. You generally sign saying the work is yours and promise to hold them harmless if sued.

---

<div class="post-metadata">

**Author:** ![Pleonast](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/pleonast/32/1183_2.png) [@Pleonast](https://boards.straightdope.com/u/Pleonast)\
**Post date:** [September 27, 2023, 3:37pm UTC](https://boards.straightdope.com/t/ai-use-opens-hollywood-to-public-domain/990549/19 "2023-09-27T15:37:42Z")

</div>

There’s going to be a lot of messy lawsuits in the future.

For example, recent Nvidia graphics cards have an [AI upscaling](https://www.nvidia.com/en-us/shield/support/shield-tv/ai-upscaling/) feature, where what you see on the screen has been modified by AI. Does this mean that screenshots of a game using this filter are now in the public domain? That feels wrong to me, but I’m not sure how these rulings would make it otherwise.

So it seems to me that _some_ amount of AI in artwork doesn’t prevent it from being copyrighted. And that “some” will be litigated for a long while.

---

<div class="post-metadata">

**Author:** ![DrDeth](https://avatars.discourse-cdn.com/v4/letter/d/b487fb/32.png) [@DrDeth](https://boards.straightdope.com/u/DrDeth)\
**Post date:** [September 27, 2023, 9:02pm UTC](https://boards.straightdope.com/t/ai-use-opens-hollywood-to-public-domain/990549/20 "2023-09-27T21:02:16Z")

</div>

> [@Pleonast](#):
>
> For example, recent Nvidia graphics cards have an [AI upscaling](https://www.nvidia.com/en-us/shield/support/shield-tv/ai-upscaling/) feature, where what you see on the screen has been modified by AI. Does this mean that screenshots of a game using this filter are now in the public domain?

The ruling only included works made solely thru AI.
