“I’m increasingly inclined to think that there should be some regulatory oversight, maybe at the national and international level, just to make sure that we don’t do something very foolish. I mean with artificial intelligence we’re summoning the demon” —Elon Musk warned at MIT’s AeroAstro Centennial Symposium
Artificial Intelligence is not considered as a legal entity that is recognised by law. For instance, if we look at companies or corporations, they have a separate legal personality – this is a legal fiction created by law. If there is any malpractice, fraud etc, the corporate veil is pierced, and the court looks at the people behind the veil. Those persons will be held accountable for fraudulent acts. If the very same logic is applied to Artificial Intelligence, we can either hold the user or the software developer accountable for any act of infringement. This, however, raises disturbing questions and defences that will be raised in the following section.
In the popular monkey selfie case[1], a monkey had accidentally taken a selfie on a photographer’s camera. The question that arose before the court was, with whom does the copyright of the photo vest with – the monkey or the photographer? PETA an animal rights organization claimed that the copyright vested with the monkey. However, the photographer claimed that the copyright vested with him as it was his camera and also because he had provided creative input. The court in the case, went ahead and held that the copyright vested with the photographer and not with the animal. The rational provided by the court was that an animal cannot be a creator or author of a work. Here, we see that the court is trying to find the closest human link – we see the same rationale being used in several cases. That being the case, where does Artificial Intelligence fit in this picture?
In the case of Yogendranath Naskar v. CIT[2], the question that arose before the court was whether a Hindu deity could be assessed under Section 3 and 4 of the Income Tax Act. The court opined that “there is no reason why the meaning of the word individual in Section 3 should be restricted to human beings and not juristic entities”[3]. The court further went onto hold that they are capable of paying taxes through their representatives. If we inculcate the very same logic to Artificial Intelligence – a machine that is far more intelligent and capable than an idol, can be assigned a legal personhood.
The concept of originality and artificial Intelligence
In the case Eastern Book Company v. D B Modak[4] the court introduced the idea of ‘modicum of creativity. The court held that “The word ‘original’ does not mean that the Work must be the expression of original or inventive thought. The originality which is required relates to the expression of the thought. But the Act does not require that the expression be in an original or novel form, but that the work must not be copied from another work – that it should originate from the author; and as regards’ compilation, originality is a matter of degree depending on the amount of skill, judgment or labour that has been involved in making the compilation”[5]. The court shifted its perspective from the sweat of the brow doctrine to the modicum of creativity. There is a possibility that AI related work can be considered as original, primarily because the act of compilation of work from already existing databases could be considered as a valid use of skill and judgement. An argument that could be raised against this theory, is that it is a human that sets up the parameters within which Artificial Intelligence works. However, merely because the parameters were set up by a human it does not deem the work to be made by a human, because at the time the work was created by the AI there was no human interference.
In Nova Productions v. Mazooma Games[6](English case) the court had to decide whether the players input constituted an authorship of a computer game. The court held that the player has neither contributed in terms of skill or judgement of any kind and neither was it artistic in nature.
In regard to Artificial Intelligence, it may be said that “the question whether AI can create original work is still debatable. Under the Copyright Act 1957, literary work recognizes compilations and since the AI is dependent on the existing information and the exposure of the programming, the work so created may qualify as compilation and therefore protectable as copyright”[7].
Today, we have an entire industry (music industry) which is “built around AI services for creating music, including the aforementioned Flow Machines, IBM Watson Beat, Google Magenta’s NSynth Super, Jukedeck, Melodrive, Spotify’s Creator Technology Research Lab, and Amper Music.”[8] These systems make use of deep learning networks whereby the software is fed with a lot of material. The Artificial Intelligence system analyses the patterns from the large amount of material provided and creates its own music/melodies.
From a basic perusal of these ideas, it is clear that Artificial Intelligence generated works do fall squarely within the ambit of copyrighted works. The larger question that raises alarms and ambiguity is, with whom does the copyright vest with? Even if we deem it plausible for Artificial Intelligence to create copyrightable works, the issue of who takes accountability in cases of infringement arises.
Who takes accountability in case of infringement?
Section 51 of the Copyright Act prescribes when a copyright is infringed. On prima facie reading of the provisions, it is clear that only a person can infringe a copyrighted work. In the case of AI, whom does the liability fall on?
There are several issues that may arise due to assigning Artificial Intelligence a legal entity. For instance, if there is an infringement of an already copyrighted work – Whom does one sue? Do we sue the user or the developer? The user can very effectively claim that it is the Artificial Intelligence that produced the work, and that the user has no knowledge that it constituted an infringement. In that case the user cannot be held liable for any potential infringement. Would this constitute a valid defence? With respect to the developer, although he has put in the requisite skill in creating the software, he has in no way contributed to the final work produced by the Artificial Intelligence machine, the same question arises as to whether this would constitute a valid defence?
Economic and Moral Rights
The primary object of the copyright act is to protect the rights of the author (economic and moral). It is worthy to note that Artificial Intelligence is not capable of suing or being sued. Moreover, it has absolutely no motivation, either financial or moral, to protect its own works. Further, the existing copyright regime recognises the right of the author to claim royalty. The right of loyalty is a right that cannot be waived. Questions as to the manner in which royalty is to be distributed, what is the amount to be fixated upon and who determines the royalty, poses a plethora of problems. In the light of these considerations, if one was to grant authorship to Artificial Intelligence – would these rights provided in the copyright act be redundant? The answer to this question is unfortunately in the affirmative. However, if the person who develops the Artificial Intelligence is granted the authorship rights, then there would be an incentive to make more Artificial Intelligence machines in the future.
Working backwards, on the logic that Artificial Intelligence needs no incentive to work and further the fact that there is no expense incurred in creating such work – a plausible alternate solution would be to put out the work in public domain without attaching any cost towards it whereby the work is not owned by anyone. This, however, may act as a disincentive for developers to create more Artificial Intelligence. Further, this would be counter-productive because it would decrease the work in the public domain. In addition to this, multiple cases of infringement could potentially arise because a person can slightly modify the work produced by the Artificial Intelligence and contend that it is his own work. Moreover, it goes against the very ethos of the copyright act – which is incentivizing the author.
Conclusion – The foreseeable future and the practical way forward:
The Copyright Act, 1957 is not equipped to deal with the rights related to Artificial Intelligence. The practical way to go about the authorship issue is to either elevate Artificial Intelligence to the level of a legal status. In such a case, the rights regarding Artificial Intelligence cannot come into question. An alternative approach is to amend the Copyright Act thereby, including the works of Artificial Intelligence as a separate category or by acknowledging Artificial Intelligence as an author. In the event of any infringement, the owner of the Artificial Intelligence will be held accountable.
One way to effectively deal with the issue of authorship is by assigning the ownership of all works created by the Artificial Intelligence machine to the developer of the software. The rationale behind this, is that it is the developer that uses his skill, judgement, labour, and time in making the software. It only seems fair that he is rewarded the copyright as a result of his intellectual labour. The pre-requisite for this Artificial Intelligence machine to function effectively and to generate its own outcome, is to first develop a machine that can produce commendable results. The development of this machine is attributed to the developer. Although, the developer may not foresee all the outcomes generated by the machine – he has spent his time, energy and expense working on it. The programmer is not merely a catalyst in this situation – but he has worked from scratch to create the software. Moreover, he has created the framework within which the software should function. The developer directly or indirectly has contributed to the final product generated by the software. It only seems fair to award him with the copyright. To strengthen this argument, we back it with a legal doctrine wherein ownership is granted to a person who does not contribute in any manner whatsoever. This is known as the work made for hire doctrine where, the employer is given the ownership rights rather than the author of the actual work. This concept is encapsulated in Section 17 of the Copyright Act. Imbibing this concept to Artificial Intelligence, it can be said that the developer is assumed to be the employer and the software can be the employee. Although, making the developer the author may be like opening a pandora box of thorny issues. Suffice to say, that this solution proposes to deal with the larger question in an effective manner. The issues that may arise later can be dealt with on a case-to-case basis.
[1] Naruto et al v. David Slater, No. 16-15469 (9th Cir. 2018)
[2] Yogendranath Naskar v. CIT, AIR 1969 SC 1089
[3] Ibid
[4] Eastern Book Company v. D B Modak, (2008) 1 SCC 1
[5] Ibid
[6] Nova Productions v. Mazooma Games, [2007] ECWA Civ 219
[7] Lucy Rana & Meril Mathew Joy, ‘Artificial Intelligence and Copyright – The Authorship’< https://www.mondaq.com/india/copyright/876800/artificial-intelligence-and-copyright-the-authorship> accessed 16th April 2021
[8]Dani Deahl, ‘HOW AI GENERATED MUSIC IS CHANGING THE WAY HITS ARE MADE’ < https://www.theverge.com/2018/8/31/17777008/artificial-intelligence-taryn-southern-amper-music> accessed 16th April 2021
Very nice presentation of value.