Name of the Student: Akansha Sharma (Student, National Law University and Judicial Academy, Assam)
INTRODUCTION-
How bizarre is the fact that today, machine intelligence is probably the last invention humans would ever have to put their minds to? Believe it or not, the time has come- the advanced neural network of machine learning software no longer requires human dependency to learn1. Thanks to the fashionable knitting of the tapestry of modern technologies, knowledge is now being automated by machines leading to mass-scale augmentation of self-generated works of great calibre! While this does reflect a moment of glory, it significantly opens a huge doorway of uncanny perplexities, no reasonable man can solve at the surface. What happens when an AI-generated work is released? Is it original? Can it be copyrighted? Is there a need to change? In order to possibly understand the nuances, we need to address the technical stages involved.
Stage 1: The Dilemma- Identifying the Hornet’s Nest.
How does AI generate anything? The answer is simple; the family of learning algorithms2 is what helps the AI simultaneously process and analyze huge chunks of extant data, thereby emulating them and procreating something which it claims to be novel. This is congruous to the process of learning. Humans assimilate in a similar way- we absorb knowledge from our surrounding data and then analyze them to create something new. It then becomes attributed to us- we own our creation. This is what gives as a locus standi in claiming copyright- the feeling of “mine” kicks in and so does the fear of protecting what belongs to us. Henceforth, it is reasonable to conclude that, every work of novelty is rather an amalgamation of a series of unique adaptations and modifications of the already existing. Since the process of learning in both humans and machines has the same foundation, if the work created by the former is deemed to be original, (in cases where it surpasses the test of modicum of creativity)3, the work generated by AI should also fall under the category of original, if the same is shown to be achieved by it.
Stage 2: The Law- Analyzing the Glitch.
In situations that do not involve machines, the law is transparent- the one who has created it, owns the copyright. But what happens when the work is generated by a machine, does the right belong to the creator of the software? Or the user?
Consider when RAGHAV painted the ‘Suryast’3– it was technically the AI that made the art, however when it tried to copyright the same, the application was rejected. It was only when Ankit Sahni, (the creator of RAGHAV) filed another application wherein, he and the software both were adduced as co-authors, the copyright was granted. It is interesting to note that in India, as per the Copyright law, the authorship in cases of machine-generated work lies with the person who causes the work to be created4. A similar stance is practiced in China, when on the grounds of proven human intervention, Dreamwriter5 was copyrighted. The UK model6 also vastly follows the Indian approach. On the other hand, the US system only protects works produced by human fruits of labor under its copyright regime. Henceforth, it withdrew the copyright protection granted to the Midjourney’s comic, Zarya of the Dawn7. Again, there is the system of France8, which under its law has clarified that only natural persons can be recognized as authors under copyright law, the open-ended system of Australia9 on the other hand makes no such distinction.
In cases where AI-generated works are being granted the copyright, they often escape infringement by hiding under the legal garb of “fair-use.” Therefore, it is apparent that the in the absence of a uniform system, the decisions of the several courts are strictly on a case-to-case basis10, which furthers whips up the dilemma.
Stage 3: The Catalyst- Paving way for a sui-generis system.
It is unassailable that the world ahead belongs to Sophia11 and her companions- humanoids capable of functioning in complete isolation from their creators. The $700,000 worth-art which she sold recently12 was made “fully” by her, as she claims. The question that begs here is who owns the copyright of this art- Sophia? A humanoid who is also a recognized citizen in Saudi Arabia, or her parent creator; Hanson Robotics?13 The 20,000 pieces of Celtic music composed by Bot Dylan14 is yet another example that demonstrates the issues augmented by the discombobulating legal position.
With the hope of adding approximately USD 957 billion of boodles to the Indian economy by 203515, it is certain that the laws ahead would be favoring AI-induced works. In such a scenario, it is time that the law adapts to the inevitable. There is a need for breeding a new system that would oversee the legality of AI-related works. The WIPO too, engaging in numerous deliberations has ruled out the same16. While the major countries and the IP-related organizations agree on the same, the modalities and the modus operandi of this new system is still in the cold.
The author believes that firstly, a new law must be created. A mere modification of the present law would open new difficulties related to liability and the right of suing in the courts. Without addressing the underlying principles of locus standi, such changes risk creating ambiguities as to who has the legal capacity to initiate claims, thereby leading to inefficiencies in the judicial process. Moreover, any alteration without careful consideration of causation doctrines may result in an unwarranted expansion of legal responsibility, potentially overburdening the courts with frivolous claims.
Secondly, the proposed law must only cover AI-generated works independent from any human intervention. The ones created with the aid of humans must fall within the folds of the present system alone. This categorical differentiation would abet any obscurity regarding the issue of authorship. It would also procreate a transparent system where identification of the specific rights and liabilities of the parties involved would become plain-sailing.
Thirdly, whether we embrace it or not, the rise of AI portends imminent involvement in legal disputes, necessitating stringent legal frameworks for independently produced AI works. Under the present system, only AI-works created with the help of humans are granted protection, the rest are denied. The consequences of this are huge- not only does it leave a huge quantum of work out of the legal spectrum (increasing chances of infringement, illegal use of the money generated by the creation due to lack of monitorship and unfair competitive practices) but also denies to accept the future possibility when the world would have independent AI’s living the lives of humans. This is against the nature of law- it is meant to positively metamorphize and adapt to the society and not deny potential change and be rigid. Therefore, the proposed law must explicitly lay down the rights and the corresponding liabilities, emerging from the copyright in cases of independent AI produce, including the legal mechanism to be used and the remedies to be sought. It must in addition map out the scope of protection granted to AI-generated works, specifically addressing whether such creations would be entitled to the same level of rights as human-authored work. Such a system would require an examination of whether the conventional criteria for copyright (such as originality and human authorship) can be satisfied by the sovran AI systems. Other aspects which require detailed scrutiny is the articulation of legal obligations of AI developers and operators in cases of infringement, and the establishment of an equitable distribution of legal liabilities. It is only when the accountability is based on strict grounds (such as direct and significant creation or dissemination of infringing work) should the entities be held responsible.
CONCLUSION-
The future of copyright lies in balancing innovation with the rule of law. Taking a stroll back to the legislative intent for the need of creating a system of copyright, the purpose was clear- to motivate creativity and ingenuity by way of protection of rewards17. The waves of revolution that comes with AI is immaculate. Its capacity to independently create requires not only a microscopic reevaluation of “authorship” but also the formation of newer doctrines that honor both technological advancement and the core principles of the law of copyright. Without any clear guidelines, we risk stifling innovation or diluting the integrity of existing copyright protections. It is only through the collaborative approach of the legislators and the judiciary that we can adopt policies that are a cut above the present ones. The need of the hour is to fortify a framework which allows the harmonious co-existence of both human and machine-made works. As we chart this uncharted territory, it is finally time for the law to meet the machine on equal terms. With AI being the future, it is time that the legislators and the judiciary sip the reality potion and step up to build a system to protect both copyright and uphold the sanctity of the law.