Many books have been written and just as many films have been made trying to show what the future will look like, but we are much closer to knowing it with the advent of artificial intelligence and, more specifically, generative artificial intelligence. Generative artificial intelligence is a type of artificial intelligence that can generate new content, as opposed to systems that simply analyse or act statically on an existing data set.
- Introduction on ChatGPT
One such generative artificial intelligence system is ChatGPT which, since its launch, has captured the attention of the whole world, with social media and online space being flooded with screenshots, opinions, videos about the most interesting, weird or funny conversations people have with ChatGPT.
What follows is a brief look at the legal implications of using ChatGPT, but not before we present what it is and how it works.
- What is ChatGPT and how can it be used?
ChatGPT is a chatbot, in other words a communication interface, developed by the OpenAI artificial intelligence research lab that is capable of responding to user requests in a conversational way, with the ability to generate human-like text based on the context of the conversation. Unlike other chatbot software, which is limited to a set of predetermined responses, ChatGPT can generate responses spontaneously, allowing it to engage in dynamic and varied conversations. Thus, ChatGPT can be used for a variety of purposes, from writing a book summary or composing a song (based on a series of words provided by the user), to converting a piece of source code from one programming language to another.
- Can ChatGPT generated content be used?
Considering the diverse purposes and the ability to often generate human-like text, the utility of ChatGPT, even at prototype level for such a technology, is indisputable. However, the question arises to what extent the content generated by ChatGPT can be used legally, given that in generating responses it also uses information and documents in respect of which certain persons own intellectual property rights.
- Legal implications in terms of intellectual property rights on the content used by ChatGPT
The risk of copyright infringement
Some of the information and documents stored in the ChatGPT database and used for “training” and further generation of responses to enquiries submitted by users are subject to copyright and are protected, at least under the applicable European legislation, against any unauthorised use, such as the reproduction of the work or the making of derivative works based on the protected work.
Furthermore, given the information and documents used by ChatGPT, the generated response will, on a case-by-case basis, contain different proportions of copyrighted material, so that any use of the generated response could qualify as an infringement of the rights held by the owner of the information and documents used in this process.
Limits to the exercise of copyright and the fair use doctrine
With regard to the use of information and documents by ChatGPT, at European level, among others, the provisions of the Directive on Copyright in the Digital Single Market are relevant, which provide for a general limitation on the exercise of copyright for “reproductions and extractions of lawfully accessible works and other subject matter for the purposes of text and data mining”. In other words, copyright holders cannot object to the use of information or documents that are lawfully accessible to ChatGPT for the purpose of text and data mining. However, the exception only applies as long as the use of the works and other protected subject matter has not been expressly reserved by the copyright owners. Thus, copyright holders can effectively prohibit the reproduction and extraction of data in various ways, such as by regulating this in the terms and conditions of a website or by using a protocol to exclude robots from using the work.
Similarly for the ChatGPT-generated content hypothesis, unless an exception or limitation to the exercise of copyright applies, the user may be exposed to the consequences of copyright infringement.
B. Legal status of ChatGPT generated content
Can content generated by an artificial intelligence system be subject to copyright?
In Europe, the Court of Justice of the European Union (CJEU) has consistently held that copyright can “apply only in relation to a subject-matter which is original in the sense that it is its author’s own intellectual creation” (see in particular the CJEU decision in the landmark case Infopaq).
Thus, it has been consistently held that a work can be considered original only to the extent that it reflects the personality of its author, with the result that in order to benefit from copyright protection the work must represent the creation of a human author.
Also, in the US, the Copyright Office as well as courts in Australia have expressly stated that a work can only be copyrighted if it is created by a human being.
In another approach, as presented in the article “Artificial intelligence and copyright” - published in WIPO Magazine, there are several countries such as India, Ireland and the UK that recognise and assign copyright in computer-generated works to programmers or persons who have made the necessary arrangements to create the work.
Who owns the copyright on content generated by an artificial intelligence system?
At EU level, given the applicable legal framework and the decisions of the CJEU, in the absence of recognition of copyright in content generated by an artificial intelligence system, the issue of who owns such rights is irrelevant.
Considering technological advances and the evolution of artificial intelligence systems, it remains to be seen whether the paradigm will change with regard to works created with minimal or even no human intervention.
Until then, it is interesting to note that ChatGPT has already been credited as a co-author on a number of scientific articles, although it seems that the editors generally agree that ChatGPT does not qualify as an author but consider it necessary to point out that artificial intelligence contributed to the writing of the articles.
But what does ChatGPT say about using its own content?
According to the article “ChatGPT - legal challenges, legal opportunities” written by the lawyers at Fieldfisher, when asked if its content can be legally used for commercial purposes, ChatGPT answered rather evasively in the sense that “it is generally legal to use the content produced by ChatGPT in commercial use cases, provided that you have the appropriate permissions and licenses. [...]. If you want to use the content generated by ChatGPT for commercial purposes, you may need to obtain a license from OpenAI or the relevant rights holders, depending on the specific circumstances of your use case.”
ChatGPT’s response therefore places the entire responsibility for determining the conditions under which generated content can be lawfully used by users on the latter’s shoulders.
We would have asked it the same question to verify the answer, but at the time of writing this articles, ChatGPT is too popular to give us an answer...
According to the terms and conditions of use of OpenAI’s products and services, to the extent permitted by law, OpenAI’s rights to the content generated using, inter alia, ChatGPT, are fully assigned to the users, but OpenAI reserves the right to continue to use both the generated content and the user-entered content, inter alia, to maintain and provide its services. The same terms and conditions also expressly provide that the user is responsible for both ChatGPT generated content and user submitted content, including ensuring that such content does not violate any applicable law.
- Conclusions
ChatGPT is a clear illustration of the evolution of artificial intelligence and, despite the limitations and concerns raised about it, it is without a doubt the future of technology.
Given the speed at which these systems are developing and the fact that the way artificial intelligence is approached will define the way we live and work, it is crucial that competent bodies and legislative authorities address this issue in a responsible way that provides timely and adequate safeguards and clarity as to the applicable legal regime.
By Monica Statescu, Partner, and Marius Gheldiu, Associate, Filip & Company