Artificial intelligence in creativity: how authorship is determined and what challenges may arise

As artificial intelligence (AI) becomes more and more firmly established in the world, not only its advantages, but also additional challenges are noticed. With the help of AI, it is possible not only to optimize the work process, but also to perform the work itself, which raises additional questions, for example, who is considered the author of the work? Naturally, as the possibilities of using AI increase, this will eventually be reflected in legal regulation. However, until there is specific regulation for this tool, users should do a lot more research on what they are using.

“Works and works created with AI have already become the object of legal disputes. Although AI seemed like a distant future a few years ago, November 2022 became a turning point of sorts. After that time, almost everyone touched DI, talked to him or asked him to help with something. Eventually, some started using AI as a work tool,” says Paulius Sartatavičius, managing lawyer at CEE Attorneys.

The expert adds that one of the more famous disputes that have reached the courts of the United States and Great Britain is related to inventions created by AI. StephenThaler wanted to patent the AI-generated inventions and the AI-generated image (visual work) stating that the inventions and the visual work were generated by AI. Simply put, the author wanted to acknowledge authorship to AI. However, both the US and British courts were of the same opinion – only a person can be named on an invention application and only a person can be the author of a work. It can be argued that with these decisions, the US and Great Britain have drawn clear guidelines regarding AI and authorship of works.

For example, the Beijing court in China has indicated in one of the cases that, taking into account the complexity of the task created for the AI ​​tool, the author of such a task also acquires the author’s rights to the generated result – a visual work. This means that the formulation of the task can in principle also be considered a part of the work, but until such a treatment takes place in international documents, then it will probably only be an additional argument in the cases.

Although the existing legal framework is sufficiently comprehensive, but considering the fact that the courts interpret the law with their previous decisions, there is a lack of real unified practice. In order to speed up the process, the lack of a uniform practice could be solved by some corrections of the legislative elements. One such element is expected to be the European AI Act, which will bring more clarity.

The author of the work artificial intelligence?

According to P. Sartatavičius, the originality of the work plays a big role in solving the question of authorship, that is, the extent to which the work contains new elements compared to similar works. Therefore, first of all, you need to answer the question whether the creation created with the help of AI is an original result of creative activity. If we answer this question positively, then the result of creative activity generated with the help of AI could be treated as a work. However, the existing international jurisprudence does not allow us to unequivocally recognize or deny that the result generated by AI is a work.

The lawyer draws attention to the fact that in the traditional doctrine of copyright and in abundant court practice, only a natural person, that is, a person, is recognized as an author. Neither a legal person, nor an animal, nor any other entity can be recognized as an author. Of course, this primarily concerns the author’s non-property rights, ie the right to authorship, the right to the author’s name and the right to the inviolability of the work. There are situations when the work has an author, but other parties also have the rights to the work. Publishers, producers, agents, heirs or transferees of rights, license holders, etc. – these are those entities that are not authors, but may have the author’s property rights acquired on one or another basis – the right to use the author’s work in a certain way. The same is true with AI-generated works – if we decide that the generated result is original and meets the requirements for the work, in principle, the property rights regime of such a work can be determined in the rules of use of the AI ​​tool itself.

Regulations and cyber security

Each AI tool has its own rules. Only a user familiar with them and confirming this can use the DI tool. But let’s be honest with ourselves, they often overlook the rules and just note that they “acquainted themselves” with the rules of use. This can become an expensive lesson, especially if you have created something memorable with the help of AI and want to realize it.

“The most important advice for everyone who uses the help of AI tools to generate any content is to read the rules of use, intellectual property and privacy policy of a specific tool. As there is currently no specific regulation related to the use of AI, individual aspects related to the use of AI, such as the regime of intellectual property, security of data submission and privacy, are regulated both by general legal acts and regulations governing these areas, and by the regulation of the manufacturers of AI tools themselves. The use of AI tools can lead to various legal risks that could be avoided by studying the terms of use of such tools. For example, some tools very clearly indicate in the rules that all generated content and property copyrights to it are automatically transferred to the manufacturer of such a tool, and the user can only use the generated result to a certain extent”, says P. Sartatavičius.

Cyber ​​security and data protection issues should also be kept in mind. In many cases, AI tools learn from information provided by users, so if we upload photos, names or confidential information to the tool, we actually give this data to the tool manufacturer, and in many cases the tool learns from this data and can provide relevant information to third parties based on certain requests from them.

The possibilities of using AI mean that this tool requires, and will require, more maintenance in the future. Not restrictions, but care. Although many feared that AI would take away people’s jobs, AI itself will create jobs. Even those companies that will not use AI tools directly will need to learn to understand whether their customers or partners will be harmed by using AI tools.

The reality is that the vast majority of companies using AI tools do not know the rules of the tools and lack the expertise to assess legal and security risks. Already today, companies should prepare and implement a policy for the use of AI tools, conduct audits of the tools used, and license reviews. This is necessary in order not to suffer later due to the inability to use the generated results or even worse – to receive claims and compensation requirements from AI manufacturers, not to mention the loss of data due to insecure provision of information to AI tools.