From conceptual art to AI: On the Druet/Cattelan dispute and authorship of works made by someone other than the “author”

· copyright,authorship,artificial intelligence,Maurizio Cattelan,Daniel Druet

When it comes to contemporary art, I must admit at the outset that there are fewer artists who I find more enticing, fun and provocative than Maurizio Cattelan [see here for the recent Comedian dispute in the US].

Having disclosed this “conflict” on my side, a few days ago the Paris Court of Appeal ruled (5 juin 2024 Cour d'appel de Paris RG n° 22/14922) in a dispute seeing, on the one hand and as defendant, Cattelan himself and, on the other hand and as claimant/appellant, sculptor Daniel Druet.

The subject of the dispute? The authorship of some of Cattelan’s best-known wax sculptures.


In 1999, Cattelan contacted Druet to create 8 wax sculptures, including amongst others La Nona Ora (1999), La Rivoluzione Siamo Noi (2000), and Him (2001)

broken image

La Nona Ora

While it appears that Druet materially realized the sculptures, Cattelan came up with the concept and was responsible for their staging.

Over time, the relationship between Druet and Cattelan deteriorated (to put it mildly), until it ceased. In 2008, Druet also realized a sculpture, titled Le Coucou, portraying Cattelan as “innocently dumb”: the work was meant to denounce conceptual artists “exploiting” those responsible for the making of “their” works.

According to Druet, Cattelan had given him quite a broad artistic freedom or, to put it like Druet did, only “vague indications”. As a conceptual artist, Cattelan does not always make his own works but, to put as his lawyer did, “the fabrication of the work is secondary to the conception”.

At first instance Druet was unsuccessful, mostly due to procedural reasons. He thus appealed but, on 5 June 2024, the Paris Court of Appeal upheld the decision, confirming that Maurizio Cattelan is the author of the works materially realized by Druet.

The judgment focuses on procedural aspects and does not say much about substantive authorship issues, if not that in Article L 113-1 of the French IP Code (unsurprisingly) provides for a presumption of authorship in favour of the person indicated as the author of a work.

The elusive “author”

All the above said, from a broader perspective, this art case raises broader questions of authorship, specifically whether authorship is about the concept, its execution, or something else?

Like other foundational concepts of copyright law (think of work and originality), authorship is an elusive and altogether mysterious notion in international and EU copyright law. The Berne Convention does neither define nor describe the requirements needed to qualify one as author.

EU law is no more explicit. The only hints we find relate to who may be regarded an author: Article 1(5) of the SatCab I Directive refers to authorship of cinematographic or audiovisual works, Article 2 of the Software Directive and Article 4(1) of the Database Directive consider, respectively, computer programs and databases, and – more recently – the DSM Directive (Recitals 72 and 74) refers to this concept as entailing a natural person to determine eligibility for the application of its contract-specific provisions.

The same approach is found in national law. For example, under UK law, section 9 CDPA states that an author is “the person who creates” a work; in the US, the Compendium of Practices of the Copyright Office provides that an author must be a human and that:

the Office will not register works produced by a machine or mere mechanical process that operates randomly or automatically without any creative input or intervention from a human author. The crucial question is “whether the ‘work’ is basically one of human authorship, with the computer [or other device] merely being an assisting instrument, or whether the traditional elements of authorship in the work (literary, artistic, or musical expression or elements of selection, arrangement, etc.) were actually conceived and executed not by man but by a machine.

In sum: the statutes may say something about who an author can be, but not really about what makes one an author.

That is where case law comes in and is helpful, including the cases on joint authorship claims. From those, for example, we learn that authorship is not synonimous with penmanship. In articulating an 11-step test for joint authorship, the Court of Appeal of England and Wales relatively recently noted [IPKat here and here] that:

If copyright protection can extend to the plot of a literary work, even where the precise words of the work are not taken, then it seems to us to be logical to suppose that the skill which goes into devising the plot is properly to be regarded as part of creating the work […] [I]t can never be enough simply to ask who did the writing.

From the “settled case law” of the Court of Justice of the European Union (CJEU), we further infer that it is the person who made the “free and creative choices” conferring originality that may be regarded as the author of that work.

In all this, one thing, is clear: authorship entails control of the overall creative process, without such control having to extend to the entirety of such a process. For example, Jackson Pollock did not know exactly where the paint thrown on canvas would fall, yet the authorship of his paintings is unquestionable. Similarly, no one would seriously question the authorship of works created under the effects of mind-altering substances.

Authorship and AI

Similarly, if we move to Artificial Intelligence (AI), the conclusion must be the same. The question therefore is: how far can one push things using AI and still be regarded as the author of the resulting work?

In this sense, decisions like those of the US Copyright Office in Zarya of the Dawn [IPKat here] and the Beijing Internet Court in Li v Liu [IPKat here] are helpful. Neither suggests that purely AI-generated outputs are protectable: instead, they stress the need to reconstruct the creative process that led to the seemingly AI-generated output in order to identify the human author’s contribution and separate that from what is purely the “work” of the AI.

To further exemplify by referring to a case that has not (yet) been litigated. When I teach classes on authorship and AI, I like to present students with these AI-generated avatars of myself:

broken image

To obtain them, I downloaded an AI image creator app, selected and uploaded 20 selfies, and then chose the template (I went for the “historical” one) I would want for my avatars.

How plausible (or even fair) is it to say that I had no authorial input in such a process? The choice of the photographs was meant to obtain a certain result (which would have looked rather different if, for example, I had uploaded images of myself as a child instead as an adult). In turn, while I cannot claim authorship of, e.g., the template, I would be reluctant to accept that I own nothing, at least having regard to the parts of the output which resemble myself and have been generated precisely because of the images that I provided.

So, going back to Cattelan and Druet, a conceptual artist can well be an author even if their work has been realized by others following that artist’s directions. But what is if Druet actually enjoyed the creative freedom he claimed he had? Well, that is a matter of evidence which, unfortunately, these Paris decisions do not explore in any detail. One might say C’est la vie or simply wait for the next court case … A presto!

[Originally published on The IPKat on 16 June 2024]