Not Everything Needs Copyright: Lawyers Flip Out That Photos Taken By AI May Be Public Domain

Copyright

from the calm-down,-guys dept

You may recall the years we’ve spent over the ridiculous monkey selfie story, concerning whether or not there was a copyright in a selfie taken by a monkey (there is not) and if there is (again, there is not) whether it’s owned by the monkey (absolutely not) or the camera owner (still no). But one of the points that we raised was to remind people that not every bit of culture needs to be locked up under copyright. It’s perfectly fine to have new works enter the public domain. So much of the confusion over the whole monkey selfie thing is that so many people have this weird belief that every new piece of content simply must have a copyright. Indeed, during the PETA legal arguments in trying to claim the copyright on behalf of the monkey, they basically took it as given that a copyright existed, and felt the only fight was over who got to hold it: the camera owner or the monkey.

As we mentioned a few times throughout that ordeal, it really appeared that PETA’s lawyers at the hotshot (and formerly respectable) law firm of Irell & Manella had taken on the case to establish some credibility on the issue of non-human-generated works and copyright. There isn’t likely to be a rush of animal selfies (though there just was a pretty damn awesome penguin selfie — no one tell PETA), but there are going to be a whole bunch of questions in the very, very near future concerning copyright and works generated by artificial intelligence. If you look, there are already many, many law review articles, papers, think pieces and such on whether or not AI-generated works deserve copyright, and some of these go back decades (shout out to Pam Samuelson’s prescient 1985 paper: Allocating Ownership Rights in Computer-Generated Works).

But now many of these questions are becoming reality, and some lawyers are freaking out. Case in point: an article in Lexology recently by two Australian lawyers, John Hannebery and Lachlan Sadler, in which they seem quite disturbed about the copyright questions related to the new Clips camera from Google. In case you haven’t heard about it (and I’ll confess this article was the first I’d found out about it), Clips is a tiny camera that you “clip” somewhere while action is happening and it uses AI to try to take a bunch of good pictures. Sounds interesting enough, if it actually works.

But, as these lawyers note, it’s not clear there’s any copyright for users of the device, and there almost certainly isn’t in Australia where they practice:

Under the Australian Copyright Act, subject to certain exceptions, copyright in an artistic work is owned by the author, which, in relation to a photograph, is “the person who took the photograph”. Therefore, as simple as that, the owner of a Clip (or similar product) which takes photos by AI will not own copyright under Australian law, as they are not the person who “took” the photos.

Unfortunately for robots everywhere however, neither will the AI. As you might have noticed in the above quote, it is the person who took the photo who owns the copyright. While “person” is not defined in the Copyright Act, it is defined in the Acts Interpretation Act (which governs the interpretation of legislation), which provides that it includes an individual, body politic, or body corporate but not, by implication, a machine.

Therefore, the answer is that, under Australia law, no-one will own copyright in photos taken by AI. The photos simply will not be protected by copyright in Australia, as they do not have an “author” within the meaning of the Copyright Act. The Australian Federal Court reached a similar conclusion when it ruled that information sheets arranged by a computer program did not attract copyright protection.

A similar analysis almost certainly applies to the US and a bunch of other countries (including Spain and Germany) where the law is pretty clear that non-humans don’t get copyright. As that and other articles note, there are some countries (including New Zealand, India, Hong Kong and the UK) which have specifically updated their copyright laws to include a new form of copyright for computer generated works (it varies, but basically giving the copyright to whichever person was most involved in the process — which opens up a whole different can of worms).

But what struck me about the article by Hannebery and Sadler, is they don’t even stop to consider why we might not want every new work to be covered by copyright. It’s not even up for discussion in their piece. They just insist that the lack of copyright must be a problem and demand that Australia amend its copyright laws to fix it without ever bothering to explain why it’s a problem:

Refusing to afford computer-generated works copyright protection is likely to become more and more problematic, as artificial intelligence develops at a mind-boggling rate and we start seeing artistic works (like paintings, music, and even novels) created by machines.

Eventually, Australian lawmakers will have to address this issue. This may mean adopting an approach similar to that of the UK and New Zealand, whereby copyright ownership is granted to (most likely) the creator/owner of the computer program which authored the work. The alternate approach of granting copyright ownership to computer programs would of course be radical, but is certainly not outside the realm of possibility as technology continues to develop.

Notice how the lack of copyright is declared to be “problematic,” and the only debate, it appears, is between whether the owner of the system should get the copyright, or the programmer of the AI.

But that’s silly. As we wrote all those years ago, not everything needs copyright. Indeed, even for most of the modern world, we didn’t automatically copyright all works of creation until relatively recently. In my case, here in the US, it was still in my lifetime that we assumed most works were in the public domain and only granted copyright to the small percentage that decided to register.

It’s just in the past couple of decades — often driven by special interests who have built entire industries on sucking up copyrights and restricting competition with them — that we’ve reached a world where the idea of content without copyright is somehow “problematic.” But it’s not problematic and it shouldn’t be, and we should get past the brainwashing of the legacy copyright players, and recognize that not everything needs copyright, and AI-generated works most certainly do not.

In that article we wrote years back, there’s a quote from Sherwin Siy explaining why it’s unfortunate that the meaning of the public domain has changed so drastically in just the past few decades:

This is the definition of the public domain?things that are not protected by copyright. We?re used to thinking of the public domain as consisting of things that were in copyright and then aged out of it after a length of time, but that?s just a part of it. There?s also works created by the federal government, and things that simply can?t be protected?like ideas, methods of operation, or discoveries.

But, because legacy copyright interests have been so driven into so many people’s heads that everything must be covered by copyright, and everything must be owned, and everything must be locked down, some people seem unwilling to even consider that the world might not fall apart if some content is never under copyright. As we’ve seen in lots of areas where that’s the case, those industries often thrive and grow more rapidly than those encumbered with legacy protections in the form of copyright.

Hopefully, as more and more AI-generated content exists, we resist the urge to lump it all under an outdated 18th century concept that simply isn’t needed to create “incentives” for a computer to generate new works.

Filed Under: ai, australia, copyright, public domain