ChatGPT and the Rise of AI Chatbots: A need for legislative Development on uncharted Intellectual property territory

The advent of Artificial Intelligence has aided multiple people in enhancing their work by serving as a hub for relevant information, and in the brief period of its popularity, it has evolved so far as to develop its own original content derived from material currently available online.

ChatGPT, an AI model designed to generate human-like responses, has been trending for quite a while now. From helping students with assignments to data entry, ChatGPT has many uses. While, the powers of Artificial Intelligence software, especially like ChatGPT, Dall-E, Bard, etc. are impressive, it must be kept in mind that the replies generated by such software are not because of their own independent creativity, but rather of the synthesis and processing of data available in various data hubs and compiling it in a suitable way to present relevant data. The fact that its output is a processed version of data from various sources raises an important question of whether there is any infringement of the Intellectual Property of the sources, whose data was being collected to be presented to the user of some Artificial Intelligence Software without any explicit revenue or credits to the source.

Taking the case of ChatGPT, its developer, OpenAI states in its Terms of Use that the user receives all rights, title, and interest in and to Output to the extent authorized by law.[i] Another important aspect is mentioned here, which is the power of Artificial Intelligence to grant and further transmit inventorship or rights. When asked itself if ChatGPT is infringing copyright laws, it said, in the same Terms of Use, that it is the responsibility of the user to ensure that they are not violating anyone’s copyright.

The elements of intellectual property rights over the data which AI collects for processing and the rights over the data presented by such software present a conundrum regarding the legal validity of the collection of input which may even be from copyrighted sources and provides the rights of its output to the users subject to the law of the land. A piece of writing generated by ChatGPT, if later published or distributed, makes it unclear as to who should be provided the authorship. ChatGPT also has the ability to share personal data from its training datasets with its users. The functionality means it probably breaches most of the world’s data protection laws.

It becomes challenging to fix liability here as ChatGPT is an AI Model and S.63[ii] of the Copyright Act provides for the punishment of any person Furthermore, if literary, theatrical, musical, or creative works are original, only one individual can be called the author. However, both the ability of AI to create original work and the classification of AI as a person come under scrutiny.  The Supreme Court’s judgment in Som Prakash Rekhi v Union of India[iii] defines the qualification of a legal ‘person’ under Indian law and observed ‘personality’ to be the sole attribution of a legal person. It further clarified that such a ‘personality’ refers to an entity having the right to sue or can be sued by another entity. While the facts of this case do not relate to Intellectual Property Laws in any way, this is one of the most relevant cases in this context, and Artificial Intelligence is not capable of such rights. The rapid development and expansion of Artificial Intelligence into various fields and its diverse applications could not have been anticipated by previous judgments or legislations, which reasonably did not envisage any exemption for such a situation.

The interpretation of the same in foreign jurisdictions also becomes an important source for the understanding of the consequences of such advanced technology. In Stephen L Thaler v. Comptroller General of Patents, Design and Trademark (DABUS case)[iv], the petitioner tried to get a patent transferred in his name for the creation of his AI Device. The claim was quite simple, that the AI invented something, and Thaler, as the owner of the device should own the patent for the invention. However, this opened up intricate matters of law for interpretation dealing with whether AI could hold a patent, let alone transfer it. The decision held AI to be incapable of holding its own patent as it did not qualify as a ‘person’.

However, a great deal of data obtained by such software is based on data available on the internet, which may violate the Intellectual Property Rights of the source from which the software obtains the data. For example, asking AI software to design a promotional campaign for your organization may cause it to blend some pre-existing protected material to produce something for you. Using such material could create liability for any organization which uses this and may expose the organization to copyright infringement suit.

Additionally, ChatGPT being an advanced AI model, it is capable of producing increasingly sophisticated and creative output, it is perfectly feasible that the output is original and not based on existing resources. This raises questions about how intellectual property laws will apply to these works and who will be able to claim ownership and credit for them.

With the development of technology, it is apparent that we may soon be approaching a stage where Artificial Intelligence Software would be able to generate their own outputs independent from the already available web data however it would be extremely difficult to identify the source and whether the source was protected.

The rapid rate of development of Artificial Intelligence provides numerous avenues through which this technology could assist many people and could even eradicate the need for humans to carry out certain generic menial tasks, allowing them to focus on solving substantive issues, thereby increasing the productivity of society as a whole. New breakthroughs in this field can help lawyers in legal research and doctors with patient diagnoses, it might even soon replace the need to manually drive.

 The legal oversight on such a critical issue has not only left ChatGPT to fight multiple lawsuits but has also raised a question regarding the liability of any person who employs the output produced by ChatGPT. The inability to pinpoint the actual origin of any output presented by ChatGPT could create opacity in any claims by litigants claiming unauthorized use of their work by such a model and hinder them in making their case. The development of this technology, wherein artificial intelligence serves as an author, raises important concerns about it carrying out actions that fall in a legal grey area. While the issue is before the Courts for the parties to battle it out, an unprecedented issue such as this one should ideally be regulated by specific regulations framed by the legislature by taking the recommendations of subject matter experts to not just answer queries relating to the cases already filed but also to provide a roadmap for the further development of such software.

---

[i] ChatGPT-Terms of use, Open AI, https://openai.com/policies/terms-of-use (last visited May 10, 2023).

[ii] Copyright Act § 65 (2023).

[iii] Som Prakash Rekhi v. Union of India, 1 SCC 449 (1981).

[iv] Stephen L. Thaler v. Comptroller-General of Patents, Designs and Trademarks, EWHC 2412 (2020).