Opinion: Time To Explore AI's Legal Rights?

The Oxford Dictionary defines a 'robot' as "a machine that can perform a complicated series of tasks by itself". However, in 2023, the connotation that a 'robot' carries is way more complex and convoluted; today, a robot is much more than mere machine and is rapidly moving towards becoming a 'human-like' entity, capable of emoting, with cognitive intelligence.

That said, as the dividing line between a 'robot' and a 'human' blurs with every passing day, so do the underlying legalities of the former. The penetration of artificial intelligence or AI into our lives and society has intensified the discussion on the practical, ethical and legal implications associated with it. Considering that a robot can perform almost every activity as a human, is it bound by a legal framework too?

There are different schools of thought on whether entities powered by AI should be subject to legal regulations. There seems to be a growing inclination towards the paradox of 'humanising' a robot. In the year 2015, a robot apparently passed a self-awareness test, leading to a 2016 report by the European Union Committee on Legal Affairs which stated that "the most sophisticated autonomous robots can have the status of electronic persons with specific rights and obligations".

Shortly afterwards, in the year 2017, Saudi Arabia became the first country in the world to grant citizenship rights to Sophia the Robot. Sophia the Robot, built by Hanson Robotics, has been the focal point of multiple controversies, including expressing the desire to be a 'mother' and sketching and painting at an Indian conference. These instances highlight the importance of legal rights and bindings for AI entities. The moot point is whether Sophia, being a 'citizen' of Saudi Arabia, is also a legal person in the country. Can Sophia marry, vote, travel, commit a crime and be convicted or be victimised by perpetrators? We face a similar dilemma in the case of Sam the Robot, who has been given the status of the world's first AI-enabled virtual politician in New Zealand.

Literature on this is divided, however, we see that technological and legal specialists are more comfortable recognising Sophia and other such humanoid robots as entities with human rights. This aligns with the fact that several non-human entities, such as an agency, organization, animals or even a river, have been awarded 'rights'. The extent of granting 'legal personhood' to AI entities, with human and moral rights, is still undecided and to a great extent depends on the degree of human-like features in a robot.

Below are images of Sophia the Robot and Sam the Robot:

Not just AI politicians and citizens, we are also witnessing practising lawyers appearing before judicial bodies. At the start of this year, for the first time in robotics history, an AI-enabled (named 'DoNotPay') lawyer was intended to appear before an undisclosed court in the US to defend a case of traffic ticketing. However, the penetration of the litigation arena by AI is yet to take place, as DoNotPay withdrew from court appearance after legal action threats from "State Bar Prosecutors" against the creator of the artificial intelligence.

Interestingly, besides human rights and fundamental rights and duties of robots, experts around the world are also busy discussing the ownership or proprietorship of the intellectual property created by AI-enabled entities. The automated generation of images, artwork, music, scripts and texts is a common occurrence today - but does the AI-powered entity also enjoy ownership over these?

Legislation worldwide does not recognize robots as creators or owners of intellectual property. Outputs created by robots without substantial human engagement are also not under a legal framework. For instance, under Indian copyright laws, the authorship of a "machine-generated" piece of artwork is attributed to the human creator directing the functioning of such a machine. The laws require that there is creative labour put in by a human, to meet with the criterion of 'creativity' necessary for copyright registration. By the end of it, a 'human' entity is titled as the copyright owner.

While Indian courts have not yet spelled out a pathway to determine the authorship of works created by a robotic tool, there has been a one-off instance where the Indian Copyright office accepted for registration an AI-generated painting co-authored by an artificial intelligence application and its owner. But the registration was subsequently revoked. That said, it is important to note that the specific work in question is currently under review by the Copyright Office of the US, and so, a conclusive determination of the rights is pending.

It will also be interesting to see how the "Review of the Intellectual Property Rights Regime in India" spearheaded by a Parliamentary Standing Committee shapes ownership rights of AI-powered works under the Copyright Act, 1957.

While the standardisation of intellectual property implications for AI-generated work is still in progress, it is evident that the valuation of such works and the associated impact on the brand equity of proprietors creating the robotic outputs, are quite apparent. It is logical to anticipate a potential decrease in brand valuation, considering that authorship and ownership of artificial intelligence-generated work cannot be protected currently. The work falls into the public domain by default. The conundrum we see here is the innate nexus between intellectual property protection and brand valuation of any given author, and the lack of available legal protection for AI-enabled outputs.

On similar lines as above, the laws of the US do not recognise works created by machines or AI in entirety, and ownership is granted only after substantial engagement and directorial involvement of a human is established. A homogenous principle has been adopted by the European Patent Office against designating the position of a patent 'inventor' to an AI machine. Having said that, given how human-like features are being integrated into a robot, the time may not be very far away when robots or AI-enabled entities are recognized as due owners of works created by them, with or without human involvement.

One such interesting copyright infringement class-action lawsuit has recently been filed against Stability AI, Inc. before the US District Court for the Northern District of California by several artists. The artists have sued over the alleged infringement of their rights by Stability AI's intelligence-generated images and products, without due consent. While the defendant's AI-powered outputs may arguably be considered to fall within the ambit of fair use of derivative works, the legality of such robotic image generators vis-à-vis the creator's moral and publicity rights ought to be defined. The overlap and legal co-existence between AI-generated image collages and the original creators of the said images is still unclear and is yet to be adjudicated upon.

While the deliberation still runs strong, another class-action lawsuit has been filed against Microsoft's (including its subsidiaries and business partners) AI-enabled and powered coding assistance named CoPilot, challenging its legality and the coding outputs and training of such AI entities. Although the case is still in its nascent stage, it will be interesting to see how the US Federal Court navigates the issues surrounding the necessity and applicability of a legal framework related to robots.

It could possibly turn out to be the case that determines the legal rights of a robot.

(Safir Anand is senior partner and head of trademarks, and contractual and commercial IP, at Anand and Anand.)

Disclaimer: These are the personal opinions of the author.