National Portrait Gallery copyright row

Jeremy Benthams NPG portrait

Jeremy Bentham's NPG portrait

Several news sites have reported an interesting copyright case involving the Wikimedia Foundation and the National Portrait Gallery (NPG) in Britain. The NPG undertook a £1 million GBP digitisation exercise, and placed high-definition versions of their pictures in a database locked with technological protection measures. Derrick Coetzee, a volunteer for the Wikimedia Foundation, accessed the database, circumvented the protection, and uploaded 3,300 NPG pictures to Wikimedia Commons. The original portraits are in the public domain, so it would be a fair assumption that pictures of the originals would not have copyright either. However, the NPG disagreed, and sent a cease-and-desist letter to Mr Coatzee alleging copyright infringement, database right infringement, circumvention of technological protection measures, and breach of contract. This is an interesting legal issue for many reasons, chiefly because the legal status of pictures of public domain paintings is not clear in UK copyright law.

Each of the claims can be examined separately. The first one is the question of whether or not the pictures of paintings have copyright. There is an American authority on this very same point, Bridgeman Art Library v. Corel Corp, where Bridgeman Art Library had taken several photographs of public domain paintings, which were copied by Corel and sold in a multimedia CD-ROM. The court in second instance ruled that slavish copying of a painting through a photograph does not constitute originality, and therefore does not have copyright. While the ruling does not constitute precedent in UK jurisdictions, it could be taken into consideration by a court in the UK, as has been the case in many other instances where there are no relevant authorities. The heart of the copyright question is one of originality, does the taking of a photograph of a painting amount to an original work, and therefore should it be protected by copyright? The existing case law on originality seems to leave the question open. Francis Davey pointed me to Hyperion Records v Sawkins, where the Court of Appeals discussed the issue of originality, sweat of the brow and skill and labour in the creation of a copy of a work in the public domain.  Jacob LJ seems to hint that mere copies of a work may not constitute a new original work when he says that:

It was no different from the case of a person making a copy of another’s work (such as the copy of a painting or the enlargement of a photograph): the process might call for effort, skill and judgement, but it did not transform a copy of a work into an original work attracting a fresh copyright…”

He then goes on to establish a test for originality:

“In the end the question is one of degree – how much skill, labour and judgement in the making of the copy is that of the creator of that copy? Both individual creative input and sweat of brow may be involved and will be factors in the overall evaluation.”

Another relevant authority is Interlego v Tyco Industries.  In this case, Lord Oliver discusses precisely the point of originality. He comments:

“Take the simplest case of artistic copyright, a painting or photograph. It takes great skill, judgement and labour to produce a good copy by painting or to produce an enlarged photograph from a positive print, but no one would reasonably contend that the copy painting or enlargement was an ‘original’ artistic work in which the copier is entitled to claim copyright. Skill, labour or judgement merely in the process of copying cannot confer originality.”

This is an entirely open question as far as I am concerned, and I think that there is a strong argument to be made that digitisation of original paintings does not require enough skill and labour to warrant fresh copyright on the images, but I am willing to concede that this could go either way.

The second claim by the NPG is that Mr Coatzee has infringed their exclusive database rights. The NPG is correct to claim that the EU has a sui generis database right, but fails to mention that this right has been seriously eroded by the European Court of Justice (ECJ) in a series of cases (e.g. British Horseracing Board v William Hill). The database right subsists, according to s.13 of the Copyright and Rights in Databases Regulations 1997, “if there has been a substantial investment in obtaining, verifying or presenting the contents of the database“. The NPG has declared that it incurred in a $1 million GBP digitisation process, but interestingly, this sort of investment does not serve to warrant the database right. In British Horseracing, the ECJ commented that:

“The expression ‘investment in the obtaining of the contents of a database’ in Article 7(1) of the directive must be understood to refer to the resources used to seek out existing independent materials and collect them in the database. It does not cover the resources used for the creation of materials which make up the contents of a database”

This is a vital point that has been completely missed by the NPG. It does not matter how much was spent in taking the pictures and digitising the paintings, what matters is the investment in creating the database proper, not its contents. The NPG would have to demonstrate that it has incurred in substantial investment in creating the database in order to claim that it holds a database right over it. Even if it managed to do this, the ECJ has set the bar high as to what can be considered infringement. According to s.16 of the 1997 Database Regulations, “a person infringes database right in a database if, without the consent of the owner of the right, he extracts or re-utilises all or a substantial part of the contents of the database“. The question would be what is to be considered a substantial part of the contents of the database. The ECJ has commented that the term “substantial” has to be analysed both qualitatively and quantitatively, so there will be an analysis of the quantity of content taken in relation to the whole database, but also the relative importance of that content to the overall investment incurred in creating the database.

I would be extremely surprised if the NPG was able to get any results on their database right claim, particularly when one considers that the digitisation investment was financed with public funds.

The third claim is that of circumvention of technological protection measures. This claim is based on ss.296ZF(1) and 296ZA of the 1988 CDPA. These are compelling arguments, as these are restrictions against the breaking of technical measures designed to protect copyright works, and it seems clear that Mr Coatzee broke the protection built into the database in order to access the high resolution version of the images. However, this claim rests entirely on the pictures being subject to copyright protection, and as I have mentioned earlier, this is not straightforward in the slightest.

In my opinion, NPG’s strongest argument may be their breach of contract claim. They say:

“Clearly visible on every page of our client’s website where an image can be seen there is a “Use this Portrait” menu on which two links are clearly visible:

  • “License this image” and
  • “Use this image on your website”

If you click on either of these links the first thing that you are told is that you need permission to reproduce our client’s images. Your downloading of our client’s images and subsequent uploading of those images to the Wikipedia website has therefore been carried out in direct contravention of the clear rules and this amounts to a breach of contract.”

This is a browse-wrap agreement agreement, as the terms and conditions are located in a link at the bottom of the page, as opposed to a specific acceptance that is presented to the user in the form of a tick-box or “I Agree” button characteristic of click-wrap agreements. Click-wrap agreements have been accepted in the UK, but browse-wraps are still not fully recognised. In an email with Simon Bradshaw talking about this case, he maintained that there is not a contract because there is not acceptance. However, the U.S. and Canadian cases dealing with browse-warp have recognised that the acceptance can be formed if the user undertakes further action other than normal browsing which could be taken as acceptance. In this case, the extraction of images would definitely be a deliberate action other than mere browsing, and in my opinion would be enough to warrant acceptance of the terms and conditions of the case.

I am perfectly aware that this is a contentious issue, but I believe that the enforceability of browse-wrap agreements is of utmost importance for all sort of reasons, but it is particularly vital to people involved in the open and commons movements, such as Wikimedia Foundation. After all, Wikimedia content is released using open licences such as Creative Commons, which are by definition browse-wrap agreements. If one was to argue that the terms and conditions present in the NPG website are not binding, then neither would be the CC licences that protect Wikipedia articles, or the browse-wrap agreements that protect open source software. Furthermore, browse-wrap agreements are an important part of electronic commerce transactions, and they are used and assumed all over the industry. To disallow them would equate chaos.

To conclude, this is an interesting legal case for many reasons. I have not even gone into the many jurisdictional considerations, as Mr Coatzee is an American citizen. I believe that the NPG would have a difficult time in court, with the exception of the contractual case, but as I said, many of the legal issues are completely open in the UK. In a selfish way, I wish the case would be litigated, as it would provide us with some interesting precedent in the case of originality of copies of public domain works, and also on the issue of browse-wrap agreements.

Update: Interesting legal opinions from Simon Bradshaw and Francis Davey (who deals with the topic of originality within an inch of its life).

18 thoughts on “National Portrait Gallery copyright row

  1. Browse-wrap doesn't have to be binding for copyleft to work.

    If there are no rights to licence, you cannot licence them and you cannot predicate a licence on them.

    If there is no copyright in an image, browse-wrap cannot licence it or predicate a licence on it. So browse-wrap for public domain images is toothless.

    If there is copyright in an image you can ignore the (copyleft) licence until such time as you wish to redistribute the image or an adaptation. Unless it's illegal to download an image from any website. So you can use download the image no problem the same as any other image. When you need a licence is when you want to redistribute or adapt the image. At that point you either admit that you have a licence and agree to abide by the terms of copyleft or deny that you have a licence and break the copyright.

    The crucial factor isn't agreeing with the licence when you download the work, it's the existence of the copyright when you go to redistribute (or perform some other act restricted by copyright).

    So I don't think that click-wrap or browse-wrap is relevant to copyleft licences, and trying to make it so would have the unacceptable side-effect of allowing the public domain to be enclosed by them.

  2. Hi Rob,

    Interesting comment as usual. The NPG's terms and conditions are vague, but they seem to imply an obligation to contact them for further reuse.

    Contract can be used to protect works that are not protected by copyright. There is precedent of situations where public domain works have been restricted through contract and not through licences. For example, a lot of genetic databases contain information that has been released on the public domain, but they have restrictions on reuse. Hap Map used to operate like that; I haven't looked at their terms recently, but they used a click-wrap agreement that restricted commercial reuse of the data. The new ODbL uses precisely this construction, as databases licensed may not have copyright protection in the U.S. , so it uses contract to bind the parties.

    Again, I think that the contract claim is the strongest in NPG's case, but it depends entirely on their particular version of browse-wrap being recognised as a contract, which I grant is not certain either. We do need a browse-wrap authority in the UK though. Badly.

  3. Andres,

    I don't think it's acceptance that's the problem with the contract claim – it's consideration. In English law (and the NPG maintains that this is the law applicable) a contract requires not just agreement and an intent to create binding legal relations, but also a mutual exchange of consideration. Each side has to give something of value to the other.

    Such consideration need not be financial or expressly agreed. In the famous Carbolic Smoke Ball case, the court discussed the example of a person who put up a notice offering a reward for the return of a lost dog. Not only is acceptance implied by the act of someone looking for and finding the dog, but doing so is a burdensome task that constitutes consideration for the benefit of the dog owner. (The reward is the owner's consideration for the benefit of the successful dog-finder).

    Equally, when Wired.com used one of my CC-licensed images from Flickr, consideration flowed both ways. Wired.com got an image they would otherwise have had to pay for, and I got the biggest boost in my image-viewing stats I've ever seen. I would argue this provided mutual consideration for a contract by which we both agreed to respect the terms of the CC licence for that image.

    But in the NPG/Wikipedia case, where is the consideration from Mr Coetzee to the National Gallery? He is not publicising material that would not otherwise in some form be online, as the images are available for viewing. He is, according to the NPG, risking damage to their revenue stream. This transaction, if you want to construe it as such, benefits Mr Coetzee but not the NPG, so their claim that it is a valid contract must surely fail.

    What the NPG is really asserting, I think, is a sort of trespass on their website in that Mr Coetzee has visited it but gone beyond the agreed terms of use. This would be analogous to what would happen if he had been caught taking photographs in the NPG itself. Even though there is no admission charge, entry is on the basis that visitors are licensees whose licence to visit is conditional on complying with the NPG's rules. Break those rules – one of which is 'no photography' – and you become a trespasser. This raises the interesting question of whether English law recognises a civil tort of unauthorised computer access as distinct from the comparable criminal offence under Section 1 of the Computer Misuse Act.

  4. I'm not sure that Simon's argument about "no consideration" will run.

    Seems to me that the theory would run like this: the viewer has gained access to the pictures, in return the NPG has gained a pledge that the viewer binds themself not to use the images in particular ways.

    Is this so different from the average Non-Disclosure Agreement (NDA), where one party grants the other access to information, in return for a pledge to preserve its confidentiality?

    But it's far from clear to me that Coetzee has actually assented to any such undertaking here.

    This whole discussion is also conditional on whether English law is actually applicable here. It seems to me that if any case came to trial, Coetzee's first move would be to try to get it moved to the United States; being a case of a corporate body suing an individual, there's a good chance that would be successful.

    Even if the English courts didn't give up jurisdiction, it seems to me the NPG would have a very hard time enforcing any such judgment in a U.S. court, as Mr Coetzee's lawyers would presumably argue:

    * That the UK courts had no personal jurisdiction over Mr Coetzee.

    * That, under U.S. law, the UK court had no jurisdiction over the subject matter; and

    * That the judgment was repugnant to the public policy of the state where enforcement was sought (ie the United States).

    A U.S. court might be expected to find those arguments considerably persuasive.

  5. The question of consideration is certainly and interesting one, but the NPG's contract claim would also have to overcome s.171(2) of the CDPA:

    "Subject to those savings, no copyright or right in the nature of copyright shall subsist otherwise than by virtue of this Part or some other enactment in that behalf."

    The NPG's claim appears to be trying to enforce a copyright through contract law, something that the CDPA excludes.

  6. You are wrong about Interlego. This decision related to an attempt to extend design right protection by asserting copyright in design drawings. In any event, the principle in Interlego was expressly doubted in Sawkins v Hyperion and has been essentially limited to its facts.

    Under UK law copyright subsists in those photographs.

  7. Interlego was a Hong Kong case under the 1956 Act that was decided on different grounds, nobody is claiming that it is a binding precedent. However it was not "expressly doubted" in Sawkins v. Hyperion, it was followed! Just as it was followed in The Reject Shop Plc v. Robert Manners [1995] FSR 870. Such decisions date back to Nottage v. Jackson [1883] 11 Q. B. Div. 627…

    There is no UK copyright in the NPG images, something that the NPG should have known had it made even the most cursory examination of the relevant law.

  8. Regarding Simon's argument about consideration, while a photo on Wikipedia is not quantitatively different from a photo on NPG, it would seem to me that it is qualitatively different. By which I mean that Wikipedia is a top 10 website and the NPG site is not, so the same image on Wikipedia is far more likely to be viewed. Each of those images has an associated description page which lists NPG as the originating institution and providing a backlink to the NPG site. I would be very interested to see any effect the uploads by Coetzee have had on incoming traffic to the NPG site.

    For example, if one searches for "Francis Bacon," both the NPG site and image are buried many pages into the search results. However, the Wikipedia article for Francis Bacon is the first result, and the first image on that page is the NPG image uploaded by Coetzee. And the reason it is the lead image is because it is of a higher quality than the previous image in that location, so a Wikipedia user replaced the image on the article shortly after Coetzee's uploads. In effect, Coetzee has provided a shortcut for the NPG (web search -> Wikipedia article -> Wikipedia image description -> NPG), as compared to relying on its relatively low webrank, that has raised its profile and drove traffic to its site. While I don't know whether the financial benefits of this webtraffic at least even out the possible losses from distribution, it certainly assists the NPG's 1856 stated goal "to promote the appreciation and understanding of portraiture in all media."

    I would thus disagree strongly with Simon's conclusion: "where is the consideration from Mr Coetzee to the National Gallery? He is not publicising material that would not otherwise in some form be online, as the images are available for viewing. He is, according to the NPG, risking damage to their revenue stream. This transaction, if you want to construe it as such, benefits Mr Coetzee but not the NPG, so their claim that it is a valid contract must surely fail." Mr. Coetzee's sole gain from his work, prior to dubious fame after becoming the center of an international copyright dispute, was the appreciation of his fellow editors on the website, as he is not paid and even edits under a pseudonym. In return, he raises the profile of and drives traffic to the NPG and provides images of value to the projects of the not-for-profit Wikimedia Foundation, which coincidentally assists the NPG in its mandate. This seems to me to clearly be "something of value."

  9. @BT, apparently there is already an FOI request to obtain exactly the information you're looking for from the National Portrait Gallery.

    In the meantime, I can only quote a similar case from Norway. Galleri NOR is the image database of the Norwegian National Library (and, I think, at least one other Norwegian museum). A while back, they asked to have some images taken down from Wikimedia sites claiming copyright. I wasn't party to the discussions (nor do I have any right to speak for those who were), but they were persuaded to check the traffic statistics on their website. Result: 60% of hits on their website came directly from Wikimedia sites! Taking into account that the people coming from Wikimedia sites were more likely to look around the Galleri NOR site than those who came directly (logical, the people who come directly already know about the site, so probably know what they're looking for), they estimated that Wikimedia was responsible for 80% of their total traffic.

    The figures would be lower for the National Portrait Gallery, I'm sure, because it has a higher profile (and is based in a country with a larger population) than Galleri NOR. As you say, it will be interesting to find out what they really are.

  10. @BT: Your argument seems to be that the contract is valid because Mr Coetzee has benefited the NPG. But if the contract is valid this makes out an element of their claim against him; indeed, it makes out what is likely the strongest element of their claim if it is valid, because a contract claim is much more clear-cut than a copyright claim.

    However, the same benefit could have been given to the NPG if medium-resolution images were provided by the NPG to Wikipedia, which is what I believe the NPG has offered to do.

  11. I don't see that the benefit to the NPG mentioned by BT can possibly constitute consideration for a contract. The benefit of increased publicity for the NPG only comes because Coetzee went beyond the terms of the NPG offer: had he accepted the offer, there would be no benefit for the NPG at all. Hence the NPG's site terms cannot constitute a contract, but rather a licence. In either case, the NPG cannot claim copyright-type rights because these are preempted by the CDPA.

  12. To address two issues, circumvention and contract:

    Circumvention:

    The "zoomify" tool works like this:

    A user using zoomify to see an image, is shown the high resolution image for that image part at a time. Individually, no circumvention is needed to see those parts. Any user may save those parts when shown, and from them put together, jigsaw fashion, the entire image. There is no circumvention involved in that process, nor effort made to prevent them doing so.

    In essence, looking purely at the claim of breaching a technical means of prevention (separate from contract or copyright), the case is this: –

    John gives Jane a 20 piece jigsaw and specifies that she may see only one piece at a time, but there is no technical obstacle in the game to prevent her duplicating each piece for herself and doing as she will with it. (Dealing here not with moral, copyright or contract but purely with technical actions needed.) Can one say that Jane is breaching a "technical means of prevention" if she does just that – by copying for herself each piece (without effort or breaching), she then freely and without obstacle reassembles the same image.

    Contract:

    It would be a matter of fact whether a contract existed (actual or implied) concerning the accessing of the image pieces and then saving them locally. It was not anticipated by NPG that they would, it was believed it would be inconvenient, but was a contract in place forbidding it?

    The question of whether NPG got a benefit (that they may not have actually wished for) such as extra visitors to their website, would be a part of the decision whether to sue, not a reason there was or was not a contract. Likewise whether or not Coetzee got benefit from his actions or not, is not material. The sole question is, did he have an agreement not to do as he did. There's a complication here, in that contract law itself varies, not just copyright law.

    Assuming that visiting a UK website, Coetzee bound his visit terms to be governed by UK law — and ignoring that CDPA says no right similar to copyright may arise other than by copyright law (a separate issue) — then one might argue:

    Offer, acceptance and consideration:

    The strongest argument would be something like this: NPG offers to the world, the ability to view one or more digital images from its archives, on conditions linked to from its website. A person accepts those conditions when, notified conditions exist, they nonetheless then visit the viewing page and are provided (by NPG) a means to view the images.

    Unfortunately NPG's own letter undermines that argument, to my mind fatally. They state:

    (I have to rely on this as the image viewing system has since been removed)

    "Clearly visible on every page of our client’s website where an image can be seen there is a 'Use this Portrait' menu on which two links are clearly visible:

    'License this image' and

    'Use this image on your website'

    If you click on either of these links the first thing that you are told is that you need permission to reproduce our client’s images…."

    In other words, the wording did not say "these are conditions for access", or "by accessing this section you agree to terms" or anything like that. In fact they were not presented as terms or conditions of the offer made. They were presented as ancilliary offers on a menu titled "use this image"; that users with some kinds of interest _might_ wish to visit additional pages, for example if they wished to license them, or use them on their website. A visitor _might_ visit those links, but might not. There is no requirement stated to do so, nor is it a condition they have done so to view the image. Even a visitor who wished to view the image for their website is not mandated to do so only on certain conditions. It's presented as a pure option.

    (I can't be sure – the site may have changed – but the wording suggests an even more fatal flaw. "Clearly visible on every page… where an image can be seen". The terms of the offer must be accepted first; they cannot be retrospectively required after acceptance. If this text means that the links were presented only when the image was viewed, they cannot be terms of viewing that image.)

    Unenforceability:

    The final barrier to contract would be an argument – possibly a strong one – either that this contract, or contracts of this kind by certain kinds of body, were "not in the public interest", or the term was an unfair one (Unfair Contract Terms Act 1977 or Unfair Terms in Consumer Contracts Regulations 1999).

Leave a Reply