Posted on October 21, 2020
Authored by Aditya Krishna*
“AI” or Artificial intelligence has been developing in leaps and bounds over the past decades. With many applications emerging, and a multitude more on the horizon, it is inevitable that the legal field is bound to undergo drastic changes to cater for the changing technology and times. Most AI theorists believe this is more a question of “when” rather than “if”. These legal changes, when brought about, would not just change the way our laws function but also our understandings of their underlying rationales. One such emerging field of speculation is that of authorship and copyrights with respect to original works created by such AI systems. This article forms the first part of a three-part series analysing authorship and copyright of such AI generated works.
This article, the first part, endeavours to study the recent Chinese case of Shenzhen Tencent v. Shanghai Yinxun, in order to examine how the said ruling has contributed to the global discourse on the issue at hand. This article aims to, firstly, break down the ruling in the said case and, secondly, argue how the said ruling has led to development within the confines of the current conservative approach to authorship.
Breaking Down the Ruling
Shenzhen Tencent v. Shanghai Yinxun (“Tencent case”) was a decision of the Guangzhou Province court (China) in December 2019 which grappled with the question of copyright and authorship of AI created works. It alters the direction of the discourse set by the Beijing Internet Court’s 2019 decision in the case of Feilin v. Baidu, (where the Chinese courts chose to deny copyright protection to AI created literary work even though the work was held to meet the requirement of originality) and has been widely speculated to be the floodgate towards the recognition of Copyrights of AI created works.
As per the facts of the case, the plaintiff, Tencent, was granted a non-exclusive license of the software ‘Dreamwriter’, (an AI software capable of generating literary works) and had subsequently published an article entailing an overview of the stock market using the said software. Attached with the article was a disclaimer stating that the article was one which was automatically written by the Tencent Robot Dreamwriter. Following the publication of an identical piece by the defendants, a suit was filed by the plaintiffs alleging unfair competition and copyright infringement. Before entering into what the court held in the case, it is important to first understand how the said AI writing assistance system works.
The Dream works system creates articles in four phases, namely that of data collection and triggering, writing, reviewing, and distributing. In the first phase, the AI system analyses the data collected using its machine learning algorithm and using the triggering module, it gauges whether the said data meets the requirements for creation of an article. Based on this it moves to the next few phase where the article is created via the writing engine, reviewed using the proof-reading module and is finally distributed on various online platforms using the distribution module. It is relevant to note here that while most steps happen independent of any human intervention and all occur within what is known as the ‘black box processes’ of the AI, the selection and arrangement of the input data, selection of the article structure and a few other process are controlled and dependant on humans. Having understood this basic process, we can now move on to understand what was held in the case.
The two questions before the court were that of, (a) whether the article could constitute an original work under Chinese law, and (b) whether the copyright of the same vested with the plaintiff. With regards to the first question, the court held that the piece being an analytical write up on available data of the stock market which employed creativity and novelty possessed the required degree of creative effort to be considered original. Building on this ruling, the court went on to analyse the role the plaintiff’s team played in the creation of the article. The main aspect for consideration of the court was whether the selection and arraignment of data by the plaintiff’s team could constitute an intellectual activity which resulted in the creation of the piece. Relying on the aspect of human intervention in the process involved in the creation of the piece, the court held that the copyright for the same vested with the plaintiff. The rationale behind the same was that the court viewed the algorithm merely as a means of facilitating the creative expression of the plaintiff’s teams’ intellectual efforts and not as an independent system for creation. On the point of authorship, the court held that the plaintiff was the author as per Article 11 of the PRC Copyright Law, again premising the same on the aspect of human involvement, which they saw as essential for recognition of the same.
Arrested Development Within Conservative Notions
From a technical standpoint, while most AI systems have not yet moved into the stage of functioning completely independent of human involvement, the diminishing importance of human involvement over the years hints towards the eventual end of such a form of dependence. With that in mind, the current outlook of AI being merely tools enabling the creation of the products of human intellectual efforts, is restrictive and prevents the smooth transition towards a newer conception of authorship and copyright.
While the Tencent case has been much celebrated and has been the center of much discussion, it is important to realise that the court merely retained the notion of authorship being exclusively a human concept (which cannot be granted devoid of any form of human intervention). In such a manner, much like the previous Beijing court’s opinion (in the case of Feilin v. Baidu), the current court also reinforced the conservative approach to authorship which centralises human intervention to the creative process. Nonetheless, while sticking to the conservative notion of authorship, the court did contribute to the development of the discourse by viewing such works created by AIs as being ‘integrated intellectual creations’, arising from both the involvement of human factors and processes of such AI systems.
While it may be argued that the said case could possibly open the doors to more such AI created works gaining copyright from the involvement of even the slightest human factor in the process, the present author believes that the same would not be possible until the dominant lens of authorship is shifted away from its human centralised outlook and until the importance of human involvement is decreased. With this in mind, only time can truly tell what role the Tencent case would actually play in the development of the said discourse – whether it would be the gateway into a new theory of authorship or merely remain as yet another mirage of development.
The next piece of this series attempts to theorize the outlook of Indian law towards authorship and copyright of AI created works.
Aditya Krishna is a third-year law student from Jindal Global Law School and is currently pursuing his B.A. LL.B. (Hons.). He has a keen interest in Intellectual Property Law, Technology Law and Constitutional Law.