ChatGPT – tool of the future or threat to copyright

In the past few years, the issue of the relationship between copyright and the results of intellectual activity generated by artificial intelligence has gained significant relevance, but mainly remained in the field of theoretical research. The development of artificial intelligence is defined as a serious challenge for the law in general and for intellectual property law in particular. It is expected that this could seriously change the copyright system, as among other legal issues, it is not clear how to qualify works of science, literature and art that are created using machine learning technologies and neural networks [1].

The problem quickly broke out of academic circles into practice when OpenAI developed and released ChatGPT (generative pre-trained transformer), the latest models of which allow you to create new levels of synthetic content. In particular, ChatGPT, using data from the Internet, is "trained" to generate text based on statistical data and patterns that it finds in large volumes of other texts. In addition, in the latest version of the neural network (ChatGPT-4) it is possible to process not only text, but also interact with images, audio and video content.

The infiltration of synthetic content into scientific literature has led some scientific journals to change editorial policies. For example, the journal Science published on its website a rule that text generated using artificial intelligence, machine learning or similar algorithmic tools cannot be used in articles published in scientific journals. In addition, the program cannot be the author of a scientific article. Violation of this policy constitutes misconduct.

Science.org

In Russia, the news about how a student of one of the domestic universities created a thesis in 23 hours with the direct participation of the ChatGPT neural network caused a wide resonance.

Cnews.ru

At the moment, the Internet is replete with offers and stories about cases of this or that use of the capabilities of a neural network, the number of offers (issues for a search query) is increasing every day. It is already obvious that artificial intelligence is not a fleeting mainstream, but a tool of the future that can change the existing system of social relationships.

In terms of legal prerequisites, the problem is that, by virtue of the current wording of Art. 1257 of the Civil Code of the Russian Federation [2], the author of a work of science, literature or art can only be an individual - a citizen whose creative work it was created, and ChatGPT is trained to create works that are very difficult to distinguish from similar works created by a person's creative work. In addition, text generation does not mean copying it (and the program creates exactly generative content), i.e. from the point of view of the form, the result is assumed to be "author's" and unique. This means that the issue of direct plagiarism is virtually eliminated. Thus, formally, the creation of synthetic content does not violate the basic principles of copyright.

Another problem is related to the fact that in the learning process, the neural network uses the so-called input data - the results of intellectual activity, which, as a rule, have copyright holders. But since copyright protects the objective form of the work, in this case it cannot be said that existing legal norms were violated when creating synthetic content.

In addition to the above, the use of the controversial technology has caused other problems (for example, in terms of personal data), which leads to attempts to ban it (for example, blocking / limiting the use of the controversial neural network in Italy). In addition, tools are being created to identify texts created using artificial intelligence. However, it is logical to assume that further improvement of ChatGPT models, including the ability to self-learn, will make it easy to bypass such barriers, which will eventually lead to an endless race in circles.

Skillbox Media

Given the opportunities and widespread demand for the controversial generative technology, it seems that the most acceptable response to the challenge will be to recognize the fact that in the future a significant part of human creativity will be directly related to the participation of such technologies. At the same time, the recognition of this fact will inevitably require an early adjustment of regulatory norms in the field of intellectual property law.

In particular, in the field of copyright, the following basic questions can be identified that need to be answered:

• Can a work created in whole or in part with the help of artificial intelligence be considered the result of intellectual activity (an object of copyright or related rights);
• Who owns the rights to a work created using artificial intelligence and to what extent: the query operator (the user who set the corresponding task for the program), the owner of the rights to the program with which the synthetic work was created, the owners of the input data (right holders whose works were used for algorithm training and generation), etc.

In conclusion, a balanced approach to the use and regulation of artificial intelligence will bring much more benefits than trying to turn a blind eye to relevant technological breakthroughs. After all, as science fiction writer William Gibson said: “The future is already here. It's just that it's not evenly distributed yet.”

Sources:

1.Gurko A. Artificial intelligence and copyright: a look into the future // IS. Copyright and related rights. 2017. N 12. S. 7 - 18.

2.Civil Code of the Russian Federation (Part Four) dated December 18, 2006 N 230-FZ (as amended on 12/05/2022) // SPS "ConsultantPlus"