ChatGPT can be defined as an auto-generative chat that extracts various data and information from textbooks, newspapers, websites and different articles. After collection, the data is processed using Natural Language Processing (NLP) to make it understandable and as similar as possible to the human language. Thus, it can be concluded that ChatGPT uses AI to provide answers to the queries of the users. This text-generated AI app can be used in different fields for various purposed both private and public, although multiple questions and concerns regarding its use have been posed by various experts.
One of the major traits of these AI chatbots is their ability to generate texts and images and other forms of content that cannot be distinguished from human-generated content. This often creates questions regarding the author of the content for the copyright and whether such content is eligible for the same IP protection received by the human creators. Copyright materials created by human authors in a fixed tangible form are considered “original works of ownership”. This means that the work should be available in a physical or digital form to be protected by copyright law. In the case of AI chatbots, it is unclear whether the content created by these systems can be considered original and fixed in a tangible form and whether copyright protection can be enforced. Certain claims focused on the fact that AI is a mere tool or instrument used by a human for his work and therefore the human creator should be considered the owner of the work. While certain others argued that the AI itself should be considered the sole creator given the ability to generate content without any human intervention.
As per the statement made by ChatGPT itself, the content generated by it may or may not be protected by copyright, and it will also not be owned by the AI itself.[1] As per European and US laws, the AI cannot own a copyright as it cannot be recognized as an author and does not have the legal personality which is a requirement for owning intangible assets. Attempts made to recognize an AI machine as the sole inventor have been successful so far. Thus, ChatGPT cannot own what it writes as it is merely an artificial intelligence. Another potential issue of IP infringement by AI chatbots is the risk that they may generate content that can infringe on the IP rights as it may be duplicative of contents generated by other AIs. For example, if a chatbot generates content based on a pre-existing work with permission from the creator, it can be considered infringing.
When ChatGPT was asked about the use of copyright material during the course of training, it claimed that developers took measures to ensure that their use of copyright material was as limited and as respectful as possible. For example, the data used was taken from publicly available resources such as books and websites and any identifying information about authors or sources has been stripped out. Yet on repeating the question ChatGPT had a conflicting answer to the question. It claimed that despite the efforts of the organization to abide by the terms, there may have been instances wherein the copyrighted material has been
inadvertently included in the training data.[2] ChatGPT’s argument regarding the use of pre-processed training data that excludes all copyrightable material is essentially questionable since a vast amount of data that can be publicly accessed are either copyrighted or have creative commons licenses to them and whose scope of use is vague and not regulated.
As for the infringement of Intellectual Property Rights, as per the Open AI Terms of Use, the user shall be held “responsible” for the output.[3] Copyright focuses on determining the infringer. The user can be jointly held liable for infringing another party’s reproduction right or right to create derivative works. For example, an AI like ChatGPT trained to generate a text-based response might largely create a derivative work using copyrighted material or by excessively copying another’s work. Since the AI models rely on existing models to create the content the chances of copyright infringement are extremely high. Most AI ingests copyrightable materials for the purpose of machine learning and advancement. Although cases regarding the same are in their early stages of development in the U.S and European courts, it is possible that Open AI could be sued for the purpose of using third-party content to train the text-based ChatGPT and it is also possible that the users might be sued for the purpose of using the content generated by ChatGPT and is similar to the third-party content.[4]
Although the emergence of ChatGPT has raised multiple questions regarding ownership rights and infringement issues, one cannot deny its value as a tool for creating intellectual property. ChatGPT has been used by various users to draft patents, a section of the patent, the claims and for infringement search, resulting in varying degrees of success. Despite all the utility that the text generator AI provides, it is not advised to use the same for creating patents as the information regarding the inventions is made publicly available through the use of the AI.
To conclude, Chat GPT is a landmark in the field of innovation. A text-based AI has multiple utilities and can often provide sufficient help to the users with the wide information source that it collects. Despite all the positive achievements of the app and its utility, one cannot deny the several legal questions specific to intellectual property rights, its infringement and the confidentiality of the inputted information that the app raises. From questions regarding its copyright to the infringement of existing copyrighted sources, the AI has raised speculations and ambiguity in the minds of people regarding the rules and regulations applicable to it.
[1] Intellectual property in ChatGPT, https://intellectual-property-helpdesk.ec.europa.eu/news-events/news/intellectual-property-chatgpt-2023-02-20_en, 20 Feb 2023
[2] ChatGPT: who owns the content generated? https://www.jdsupra.com/legalnews/chatgpt-who-owns-the-content-generated-2891692/, April,7,2023
[3] Terms Of Use, https://openai.com/policies/terms-of-use, March 14, 2023
[4] Generative Artificial Intelligence and Copyright Current Issues, https://www.morganlewis.com/pubs/2023/03/generative-artificial-intelligence-and-copyright-current-issues, 23 March 2023
Anna Rose
Author
I am a second-year law student at ILS Law College, Pune. I like to read various topics and books and occasionally write when I am up to it. As a beginner myself I am still exploring the field of IPR and am hoping that I can learn and write more about the same.
Leave a Reply