An increasingly higher percentage of code in computer programs is in practice being created through tools based on so-called artificial intelligence. The code of such a computer program is therefore not “written” by a human being, but rather is the output of neural networks arising on the basis of an instruction (prompt) given by a human being. It can be expected that this trend will continue in the future.
Current copyright regulation of computer programs, which has its conceptual origins in the 1980s, cannot logically directly reflect this development. Its concept is based on the protection of literary works and thus on the premise that computer programs always have their “human” author. Protection is then primarily afforded to the copyright of such person (the programmer). The existing models of contractual relationships in the development and distribution of software (licence or sublicence agreements) then proceed from this conceptual basis, which also applies to the entire area of Open Source and free software. However, the possibility of creating computer programs through AI tools fundamentally changes established practice from a legal perspective.
As has already been indicated above, under the current legal regulation in force in the Czech Republic (and also in a number of other countries), only a natural person can be the author of a computer program. This also follows from the provision of Section 2(2) of the Copyright Act (autorský zákon), which states that a computer program has the same protection as a copyright work if it is original in the sense that it is “the author’s own intellectual creation.” Copyright therefore does not protect those computer programs which were not created (written) by a natural person, because such programs are not the author’s own intellectual creation. In this context, it is appropriate to mention that legal regulation has never protected the mere functionality of computer programs. From a legal perspective, it is therefore entirely lawful to programme one’s own new text editor or spreadsheet co-processor which will have the same or similar functionality as existing software products. This will not constitute an infringement of the copyrights of entities which have previously created and brought to market similar programs.
Naturally, individual cases in practice will not be black and white, because situations where a programmer modifies code generated through AI (and vice versa) or where he has some less important functions completed still predominate. It may not be easy to determine unequivocally from what point code is the author’s own intellectual creation and when it is not yet so, and the boundary of copyright (non-)protection may therefore not always be entirely clear from a practical perspective. Connected with this is also the procedural question, namely the possibility of proving at all that a computer program was created by a human being or conversely created by an AI tool.
Even though the role of a human being may continue to be irreplaceable in software development, this is not decisive from a copyright perspective, provided that such person does not directly “programme”. Even if someone in software development gives instructions, brims with excellent ideas or devises brilliant prompts, current copyright law does not attach relevance to this. This also applies in those cases where it concerns the development of computer programs which are subject to copyright protection. The provision of Section 8(2) of the Copyright Act mentions this expressly: “a co-author is not one who contributed to the creation of the work only by providing assistance or advice of a technical, administrative or professional nature or by providing documentary or technical material, or who merely gave an impetus for the creation of the work.” The outputs of such person’s activity will therefore usually not be protected by copyright at all, which follows from the provision of Section 2(6) of the Copyright Act: “a work…is not in particular the idea for the work in itself…an idea, procedure, principle, method, discovery, …mathematical and similar formula, statistical graph and similar object in itself…”. Whilst copyright protection of the actual content of a prompt cannot theoretically be ruled out if it were the unique result of creative activity (a very complex or poetic prompt), this does not affect the above. That is, that a computer program generated on the basis of the most sophisticated prompt will not be protected by copyright at all.
In our view, it is currently in the interests of software companies to assert that all code was produced by their employees. From their perspective and from the perspective of software creators, copyright protection of computer programs is in fact very advantageous. Copyright is a so-called absolute right, which means that, similarly to proprietary rights, it operates against everyone (erga omnes). The author has the exclusive right to profit from the granting of licences to the computer program, and a computer program created by a human being therefore cannot from a legal perspective be copied by anyone without the author’s consent.
However, as mentioned above, in the case of machine-generated computer programs, this absolute protection by copyright does not exist. If a third party thus copies a computer program generated through AI (even for commercial purposes), this will not constitute an infringement of copyright. The answer to the question of whether in such a case there may be an infringement of another right of the “originator” of the software (breach of trade secrets, unfair competition, breach of an assumed contractual obligation) will be more complex, and may also depend on the factual circumstances of the case. However, the legal position of the “originator” of the program will certainly always be weaker than is the case with copyright protection.
As outlined above, a situation where a computer program is not subject to copyright protection has important impacts on contractual relationships related to such program. This also applies to the area of Open Source and free software, which is based precisely on the existence of copyright protection. Specifically on the fact that the holder of rights to a particular computer program by the provisions of the licence agreement whilst granting licence acquirers certain rights, also imposes upon them certain obligations. Compliance with these obligations together with copyright protection then ensures that the program will continue to be distributed under the licence terms of Open Source or free software. Breach of these obligations is in fact not only a breach of the licence agreement, but also an infringement of the copyright of the rights holder. This model can logically no longer function in respect of those computer programs which are not subject to copyright protection. From a legal perspective, it will therefore be necessary to reconsider all current development and distribution models and adapt them to the changed situation. This naturally also applies to the sphere of ordinary “commercial” (so-called proprietary) computer programs.
The article was originally published on the LUPA server.
This text was translated from Czech to English using an AI translator.