Can a robot write a symphony: copyright in the world of artificial intelligence
Free use of artificial intelligence (AI) has recently moved from the status of science fiction to real life: anyone can turn to existing platforms like Claude 2, Kaiber, Stable Diffusion, Beatoven or ChatGPT to create text, images, video or music. AI results are also used in commercial and creative activities, for example, to write code, create images for promotional materials, write literary works - in a few seconds the user gets a result that he himself would have done (if he could have done it at all) for several hours.
However, who owns this free and fast-generated cheese? Below, our lawyers Kamal Tserakhau and Ekaterina Erohovec will tell you more about this in detail.
Looking for a human! Preferably an author!
To understand the answer to the question of ownership of rights, you need to know the very idea behind legal protection. The object of copyright is protected not because the picture is beautiful or the book is interesting, but because they are the creative result of the author's intellectual labour, that is, a person. Different countries may have different levels of requirements for the creative nature, but the requirement for human author participation in the creation of a work is applied universally.
In addition, it should be taken into account that copyright can only apply to an object that is original, that is, the author's own intellectual creation (EU, Infopaq C-5/08). Originality, independent creation plus a certain amount of creativity are a prerequisite for protection (USA, 17 U.S. Code § 102, FEIST,499 U.S. 340 (1991). Copyright applies to original works (UK, section 1 Copyright, Designs and Patents Act 1988) where the word "original" in this case refers not to the originality of ideas, but to the expression of thought, i.e. the work must originate from the author (University of London Press Ltd v University Tutorial Press Ltd).
The Terms of Use will reveal the truth to us!
Indeed, the service usage rules usually contain a clause transferring all rights to the user. For example, OpenAI writes:
| "(a) Your Content. [...] As between the parties and to the extent permitted by applicable law, you own all Input. [...], OpenAI hereby assigns to you all its right, title and interest in and to Output. This means you can use Content for any purpose, including commercial purposes such as sale or publication, if you comply with these Terms. [...]" |
Midjourney states similarly in its Terms of Service, but with nuances:
|
"[...] You own all Assets You create with the Services, provided they were created in accordance with this Agreement. [...] Your ownership of the Assets you created persists even if in subsequent months You downgrade or cancel Your membership. However, You do not own the Assets if You fall under the exceptions below. If You are an employee or owner of a company with more than $1,000,000 USD a year in gross revenue and You are using the Services on behalf of Your employer, You must purchase a “Pro” or "Mega" membership for every individual accessing the Services on Your behalf in order to own Assets You create. If You are not sure whether Your use qualifies as on behalf of Your employer, please assume it does." |
Unfortunately, relying on the AI’s terms of use regarding ownership of the AI output will not help. The reason is simple: the AI owner itself does not acquire rights to the created object, and since nothing arises, there is nothing for it to transfer to you. This is due to the fact that the object is generated by the AI itself, without the participation of its owner's employees. This means that the criterion of mandatory creative human participation described above is not met.
Therefore, the condition of the AI's terms of use that all rights are transferred to the user is not an unambiguous legal basis for acquiring rights to the object. This is analogous to art created using a pirated copy of Adobe Photoshop, which still belongs to the author, however, does not relieve him of liability for the illegal use of the software.
So who acquires the rights?
Thus, for the emergence of an object of copyright, a human author is needed who will demonstrate his creative nature, so there can be two answers to the question of ownership of the result:
- no one is the author and rightholder, since there is no creative element of human intellectual labour;
- the AI user is the author and rightholder, since he has shown a sufficient level of contribution to the end result.
Accordingly, the question arises as to what kind of contribution the user must make in order to acquire rights to the result?
As existing practice already shows, even in the US, where creativity requirements are considered lower than in the EU, AI-generated objects are still denied legal protection.
The US Copyright Office clearly requires substantial human involvement in the resulting output. What constitutes "significant" participation is not defined, but simply typing a prompt and hitting "generate" will certainly not be enough. However, legal protection will exist if the author has made a creative contribution through subsequent processing of the result, or combined several AI results to produce a new work.
An illustrative example of when the author does not acquire rights: re-applying for registration of "Théâtre D'opéra Spatial" (recall that registration of a copyrighted work is not required in the US, but, unlike the EU and UK, such registration is needed for judicial protection of rights to the object).
The author (let's call him that) used Midjourney to create the initial image, then processed it independently in Adobe Photoshop and increased the image size in Gigapixel AI.
However, this was not enough, and the USCO concluded that the result remained predominantly generated by Midjourney, the manual change was minimal, so registration was denied.
Thus, the boundary can be drawn as follows:
- legal protection will exist if there was subsequent "manual" processing of the AI result, while simply removing some elements or increasing sharpness is not enough - the contribution must be more significant (for example, changing the colour palette used, arrangement of objects, if it is an image, changing the style of the work, etc.);
- legal protection will also exist if the result is a creative compilation of AI results or results from multiple AIs, clips created using AI can be cited as an example: Lost Linkin park or Kupalinka Naviband.
A more complex issue concerns situations where the author's contribution is limited to a prompt. Prompts vary in complexity: from two words to several paragraphs, which significantly affects the result. It is difficult to draw a clear line here regarding the "emergence of copyright", but we will discuss this in more detail in one of the next materials.
In summary, copyright in the AI result can arise either for the author or for no one. For it to arise for the author, he must demonstrate a significant degree of creativity in creating, processing or compiling the object.
Dear journalists, the use of materials from the REVERA website in publications is only possible after our written permission.
For materials coordination, please contact by e-mail: i.antonova@revera.legal or Telegram: https://t.me/PR_revera
