Court of Appeal ruling will prevent UK museums from charging reproduction fees | Hacker News Reader

Court of Appeal ruling will prevent UK museums from charging reproduction fees

theartnewspaper.com

194 pointsedent55 comments

This was settled years ago in the US by Bridgeman vs Corel (Corel could copy old photos for their library disc). Also see Meshworks vs Toyota (3D scan of real car is not a new work). It's all based on Feist vs Rural Telephone (telephone directories not works of authorship). Museums have been whining about this for years, but not winning in court. Wikimedia puts up images under those rules, and no one has ever successfully sued Wikimedia over that, although there's been some huffing and puffing.

Feist is why the US doesn't have database copyright, and that's a Supreme Court decision based on the copyright/patent clause in the Constitution. It's a very strong holding, in that says Congress can't enact laws to make copyright cover things that don't have human authorship. "The standard of originality is low, but it does exist."

Bridgeman Art Library (the company) has been arguing for years that UK law lets them own copyright of pictures of old masters. That's over.

This has implications for large language models.

I would think the transformativeness defense used by appropriation artists might be applicable to establish that a photo of an out-of-copyright image is an “original work of authorship.” Cue Richard Prince to put it to the test…

TFA explicitly addresses that was the question under consideration:

> In his ruling (THJ v Sheridan, 2023), Lord Justice Arnold wrote that, for copyright to arise: “What is required is that the author was able to express their creative abilities in the production of the work by making free and creative choices so as to stamp the work created with their personal touch”. Importantly, he went on: “This criterion is not satisfied where the content of the work is dictated by technical considerations, rules or other constraints which leave no room for creative freedom”. In other words, if the aim of a museum photograph is to accurately reproduce a painting (which it must be), then it cannot acquire copyright.

I.e. these museums were selling pictures of artworks that by definition needed to adhere to a true reproduction of the original as closely as possible. It wasn't a question of taking a picture of an out-of-copyright painting and applying some level of skill/artistic license to transform the original in some way.

It is my understanding that in collage works and the like, the artist has copyright over their creative inputs to the resulting piece, but does not gain any rights over the bits they cobbled the final together from. If I understand it correctly, even if the museum created a fancy collage for the side of the building, yes they would own that arrangement, but not any of the paintings which went into it. They still couldn't stop you from sharing images of the original works therein.

If anyone is interested in copyright law I would recommend a free CopyrightX course from Harvard ( https://cyber.harvard.edu/teaching/copyrightx ). During the course you gain insight into the legislative side of copyright matters.

Do you know if this ruling applies to word lists of the English language such as those used in tournament Scrabble?

As loathe as I am to say anything nice about our copyright system or scrabble (a territory capturing game without the most obscure and silly rules for capturing space), it does seem like the sort of thing that ought to be covered.

That may have been true in the past, but the game is now over 75 years old. There is just absolutely zero chance that the inventors of the game would have been more motivated to invent if they were given a longer exclusive term. Since the stated goal of copyright is to advance the progress of the arts, there is no justifiable reason the public should continue granting exclusivity to the game at this point in time.

The word lists are frequently updated to cover innovations in language. E.g. you'll be allowed to play VAX and ZONKEY starting in 2024.

The game itself, sure.

The list, I’m not sure. I mean if someone is actively curating a tournament list, that seems like ongoing working that ought to be protectable somehow.

Then again, various sports maintain a tournament scene without protected rules, so I may be wrong about this.

If scrabble world lists are copyrightable then why wouldn't telephone directories be copyrightable?

I believe it was settled long ago that mere lists (e.g., telephone directories) with no creative content are not copyrightable in the United States. I'm not familiar with the Scrabble word lists, but if that's all they are I'd guess they'd fall under that ruling.

Dictionaries, on the other hand, are copyrightable because there's creative work involved in writing the definitions.

I feel like there would be a case to answer if it went to court in the UK, Scrabble words are not an entirely arbitrary selection, nor are they all words in [most] regular dictionaries AFAICT. The list has been curated in order to provide a specific set of rules for a game: there seems to be a creative act of compilation. To say it another way; it's a factual presentation - but the fact (of which words are in the list) was created for the game.

This makes it unlike an almanac, say, where the list is dictated by past events.

It's perhaps similar to the Guinness Book of Records, where the choice of records to present is creative (along with the usual creative aspects of a book, composition of pages, choice of accompanying images, etc.).

One consideration a court might take is that the list cannot be generated without copying from the established Scrabble Dictionary because one otherwise would not know whether to include or exclude any particular word. A list of 'all existing English words of 1-3 letters', say, would be fine though.

I hope this is true for videos as well, in practice.

I’m worried the anti-circumvention clauses of the DMCA (in the US) or restoring / colorizing / denoising / modifying the original will be used as a loophole.

For instance, Disney+ likes to censor and otherwise mutilate content. If the rip comes from them, will they successfully argue copyright infringement on old works?

> I hope this is true for videos as well, in practice.

Videos and music is different from photos because members of production team get individual copyrights.

But I think the question is whether you can acquire new copyright to a reproduction of the film after those original copyrights have expired.

colorizing seems like real creative work (if done by humans), and restoring can be too. Denoising, IMHO, not so much, and just cutting wouldn't be either.

Manual coloring is creative. "Restoring" is a bit of an ambiguous term, but generally, it would not be original expression either. Remember, the test for originality has nothing to do with effort. Restoring something to its original state is not new creative expression, but presenting the existing creative expression.

Applying DRM to a public domain (or uncopyrightable[0]) work does not make it illegal to remove the DRM. Copy protection only has the force of law when there's an underlying copyright to enforce. There's a bit of room for shenanigans if you use a common DRM scheme that has copyrighted works protected by it, which means you can't release tools that deprotect those works, but that doesn't make the museums able to get extra copyright out of them. It is at best a speed bump.

Modifications to public domain work only accrue copyright insamuch as those modifications are creative. Chucking a public domain cartoon through a denoising filter wouldn't make the denoised version copyrighted. Colorization might accrue some copyright, but it'd be thin - basically only the exact same coloring decisions would be infringing.

[0] https://en.wikipedia.org/wiki/Chamberlain_Group,_Inc._v._Sky....

Web sites displaying non-copyrighted works could still be protected under the UK implementation of the EU database directive. There might also be non-copyright reasons why you are not allowed to reuse the displayed works without permission. Just because things are digital doesn't mean you have to comply only with copyright when making a reproduction.

It gets even better. This seems to be the court judgement under discussion:

https://assets.caselaw.nationalarchives.gov.uk/ewca/civ/2023...

I don't want to try to summarize the judgement here, but it clearly was not about museums selling licenses for reproductions of out-of-copyright works. In fact, the judgment declared that copyright infringement had taken place. I have no idea why people use this judgment to further their cause. It seems a poor fit both in subject matter and outcome.

"Surely museums will now change their policies?"

Ha, hardly! All they'll do is make low resolution photos available and restrict anyone other than themselves from taking photos (or gaining access to do so) on grounds that the images may be damaged, will fade etc.

We now need laws to ensure the photos are accessible and that reproductions are in full resolution state-of-the-art quality.

You can ban flash photography (which does cause fading) but any attempt to ban regular photography depends on superstition, not science.

Right, it makes sense to ban flash photography, but that doesn't preclude a museum making high resolution photos available for free. Also, museums will often ban photography on grounds that it can reduce or obstruct visitor egress (which is often a reasonable reason depending on circumstance).

Moreover, a museum-generated photo made under ideal conditions is going to be of much higher quality than any visitor can do on the spur of the moment.

One often sees the problem in some Wikipedia photos, that's when the only photos publicly available are ones made by visitors who shoot under sub-ideal conditions with the object under glass or in a glass case, etc. For example, years ago I tried to take a photo of the Mona Lisa in the Louvre and gave up. It's so well protected one may as well look at an official photo rather than just view it—let alone bother to attempt to photograph it.

Incidentally, that wasn't the case with another work of art—Caravaggio's The Fortune Teller. During the same visit I photographed it without flash and considering it's a pretty good photograph (it was well lit): https://en.m.wikipedia.org/wiki/The_Fortune_Teller_(Caravagg....

From rough memory the Louvre tries that "we took a photo of it, so its a new copyright" bullshit with the Mona Lisa.

Wonder if a French court will clarify the matter at some point, like this UK court did?

Do we have a shortage of free, science quality photographic depictions of the Mona Lisa?

I hope not, but the principle of the thing still matters. Even with the case in the article out of the way (no copyright on intentionally maximally accurate photographs of public domain works), the Louvre (or is it really French copyright law in general?) restricts certain non-commercial uses (including taking your own photographs) of public domain works and requires permission for commercial uses (section 4.1.1 [1]):

> a. Re-use free of charge

> The downloading and re-use of medium-format photographs published on the collections website representing works that are not protected by copyright (hereinafter called the “Photographs”) are permitted, free of charge, for any non-collective use within a strictly private context and for the following exhaustively-listed museographic, scientific and educational purposes:

[omitted]

> b. Re-use against payment

> The use for any purpose other than those exhaustively listed in article 4.1.1 a. above, and particularly any commecial use such as the manufacture and distribution of derivative products, audiovisual and multimedia production and printed publications other than those referred to in article 4.1.1, must be the subject of a written request sent by the User to Rmn-GP via the website of its photography agency, photo.rmn.fr, or by email to agence_photo@rmngp.fr.

[1] https://collections.louvre.fr/en/page/cgu

It's curious that there is no mention of this ruling's possible effect on AI art. A side-effect of the ruling may be that AI art is uncopyrightable in the UK.

A law was passed in the UK in 1988 that said computer-generated art could be copyrighted by the person who set up the computer to produce it, even if the person played no role in the generation of the work.

However, in 2010 the UK (as part of the EU) signed on to the WIPO Copyright Treaty, which only allows copyright protection for works directly produced by humans.

The Court of Justice of the European Union has made rulings that machine-generated works are not eligible for copyright, nor are any works generated within a strict framework of rules that limit human creativity. For example, somebody tried to copyright a football game, the court said that even though humans were exercising creativity within the game, the strict framework of the rules of football within which the creativity was exercised made the game ineligible for copyright.

But the UK is no longer part of the EU, so European court rulings are presumably void there (although the UK is still a signatory to the WIPO Copyright Treaty). So I find it interesting that the judge in this ruling has largely reiterated the reasoning of the European court in the matter of the constraints the WCT imposes on machine-generated copyright (“This criterion is not satisfied where the content of the work is dictated by technical considerations, rules or other constraints which leave no room for creative freedom”.)

So on its face this ruling seems to nullify the 1988 law and make AI-generated art uncopyrightable in the UK. It will be interesting to see if anyone tries to establish this explicitly.

According to the following article:

https://www.vennershipley.com/insights-events/originality-in....

the Judge in the case explicitly referenced decisions of the Court of Justice of the European Union, and suggests that those decisions are still in force in the UK, despite the UK no longer being part of the EU.

The REUL act to cancel Retained EU Laws (REULs) starts to bite as of 1 January 2024, AFAIAA.

It maybe that copyright is not currently for the chop, but it seems - in my personal opinion - that the Tory government wish to cancel many EU laws in order to prevent Labour from moving us back towards the EU in the next couple of parliamentary terms. The REUL Act came from Jakob Rees-Mogg MP and it's very much in meeting with his ethos to 'salt the land' and cause damage for generations if he can get a couple of bags of silver out of the deal.

Long story short: we're still following EU law in a lot of areas, but not for much longer. As with the pandemic, the Tory government are no doubt using this as an opportunity to do nefarious deals in order to steal from the taxpayer.

In reality all generative art (both AI generated and old fashioned procedurally generated art) involves a human tweaking the knobs a lot and then selecting the best ones out of a sea of duds. To me it is clear that there is plenty of space for human creativity.

Yes, but, it will be very dangerous for somebody using AI generated art to pursue a claim because they may find that the other party generated the same image first. :)

That raises an interesting situation as I expect that one would be able to copyright the 'prompt' but not the generated image, and since the generated image does not include a representation of the prompt then someone else could generate the same image without violating the copyright since they are not reproducing/publishing the prompt. Prompts are then best protected as a trade secret.

In many ways its no different than electronic musicians building songs based on short samples of older songs. I have no doubt there will be rulings supporting the sort of creativity you're discussing at some point in the not so distant future.

But the key is that not all creativity can be copyrighted. Both US law and the Berne Convention (which almost all countries are party to) permit copyright only for artifacts fixed in a tangible medium, and it must be a human that does the fixing.

Tweaking knobs or writing prompts before the actual fixing is done doesn't count. Selection can't be copyrighted, just as ideas and mathematical formulas can't be.

How does a neural net differ from photoshop in terms of what the computer did? Both are machines being run by humans towards a goal.

I will dodge the question of whether the weights contain pirated material or indeed if our human memories do!

The headline is misleading.

This ruling does not prevent a museum charging fees for giving you a high res photo or digital copy of their ancient out of copyright artwork.

However, if they do, they cannot prevent you copying it and giving it to all your friends for free.

I doubt this will affect their business much - nobody wants to pony up the cash to buy a bunch of official photos to resell for cheaper at a competing store, even if they legally could.

(in the UK)

The US has had similar precdent for a while.

Why is that good?

It's funny because a few years ago everyone was stating this. Now that GPT and transformer technology exists, abolishing copyright is suddenly no longer popular. It's almost like this is just one giant echo chamber and things swish and swirl like the ocean on opinion.

I don't completely agree with the parent but it's definitely too long. Life + 70 so in average 110 years coverage is ridiculous. Why is that not good? Because it push studio to milk their cash cow forever (as we see with the big franchises occupying all multimedia space and monopolizing all large investments) and it just limits creativity. If the duration was reduced it would force them to invest on creators to find new successes and that would benefit the general public and the creators.

Fair use is also far too restrictive as we see with LLMs. What do you think about when you read "video game plumber"? So basically it became a foundational concept cognitively for you and me (polite way to say that we've been brainwashed by millions of $ in investment in publicity such that the concept is deeply ingrained in our minds) and on top of occupying a significant amount of space in our brains we should be forbidden to use the concept in our communication and creative work?

I don't think the pattern you're describing exists.

I've seen a few people want to abolish copyright, but not many. The vast majority always seemed to want reductions in length. And historically a lot of people have wanted fair use to be easier.

When I look at the arguments against AI, they're not asking for copyright to be longer. They're asking for it to be stricter in a very particular way. And they're making a distinction between human use and transformer use.

There's no "swish and swirl" that I have seen. There's not really a conflict between those opinions on fair use, and there's definitely no conflict with the opinion on duration.

The solution to something that has a lot of problems is not going to the opposite extreme.

being "the opposite extreme" doesn't mean it's wrong

It usually is though.

Well, it can use some rework.

A piece of technology based on theft is useless. If anything, copyright has become more relevant than ever.

I believe, if anything, copyright is more useless than ever. It's an anachronism when these powerful tools of media creation are so available to so many people. Is it more valuable to society that a select few benefit from decades of cultural appropriation, or is it better for society that cultural artifacts from my youth to be freely available to everyone? As it stands, I may even die before material from before my birth becomes public domain. That's greedy, stupidly shortsighted, and an inefficient allocation of resources.

The issue though is that art is not stolen from a select few, but from society by a select few. The roles are inverted. There’s no benefit for society when a few greedy corporations steal art they then monetise. Said “resources” are not dug out of the ground. They the fruit of people’s labour. Stealing it does no good.

If it was actually theft, I'd agree, but an idea is not able to be stolen. And who inspired the person to create? Did they steal by observing the world around them? No one is an island, and who's to say they'd even have the creative muse without the support of language and other art. Calling "copying" stealing does no good.

> “What is required is that the author was able to express their creative abilities in the production of the work by making free and creative choices so as to stamp the work created with their personal touch”.

If I read this right, if I crank Instagram filters to 11 then it can be considered as enoigh free and creative choices?

What if I only up them to 1?

Just imagine if all learning materials were freely available to everyone worldwide, instead of being held hostage by money and power.

Are articles, older than a certain age, in newspapers such as NYTimes, copyrightable?

Could have major implications for the future of AI.

Anything that was first published 1927 or earlier in the U.S. is not copyrightable. The cutoff will soon roll over to 1928.

In the US, NYTimes editions from 1927 or earlier (1928 in two days time) are out of copyright and enter the public domain.

Anything before the start of the Great Depression.

Cultural stagnation at its finest.

Reply on news.ycombinator.com