This article is written by Pratyusha Ganesh and Vishruti Chauhan, students of Symbiosis Law School, Hyderabad.

Introduction

Artificial Intelligence (AI) is growing at an exponential rate throughout the world. This boom raises the question of IP management in AI. There have been discussions and moderations but not a conclusion on the subject matter. The question as to the granting of special status to the work produced by an AI still persists. There are a number of anomalies when it comes to the regulation of IPR within artificial intelligence. There are questions with reference to the ownership of patent and copyright and great concerns over the infringement issues and the penalties involved. Even with international agreements and conventions in place, there is no clarity on the law with the advancing technology.

Countries like the USA and the UK have made persistent efforts to interpret the existing laws in a manner that encompasses the AI feature as well. The copyright issue in the case of selfie-taking monkey in the USA and the landmark Infopaq Case in Europe interpreted the existing laws and the use of words ‘authorship’ and ‘intellectual creation’ to analyse and come to a conclusion for the ownership of IPR. However, considering the rate at which AI is being used in major applications, it is crucial that proper laws are put in place. This can also mean to make significant changes in the TRIPS agreement.

Sophia– A humanoid robot was granted citizenship by Saudi Arabia in 2017.Alphago and Zero are also some of the astonishing development in the technology. These surprising developments have been possible due to the features of Artificial Intelligence (AI). The term Artificial Intelligence was coined by John McCarthy at a conference in 1956. The term was used for the ability of the computers to take decision by itself without human interference with the use of algorithms and commands. It was thought by early thinkers that a time will come when the machines will fail human intellect.

The invention of computers had strength this view and today there are actual machines and robots learning to make out the best of the human learning. AI became sensational when it was first started and within years, it has become one of the most promising features of technology. The machine learning process includes analysis of data, identifying patterns of user’s preference and applying it to get optimum result in the market. It is the high demand of manipulating and organizing large amount of data that call for the usefulness of AI.

It improves the use of the already existing applications to be used at an optimum level. For example, Siri was used to the new generation Apple products to enhance its quality and productivity. Roomba 980 is a model of AI which is used for cleaning, Google’s AI-powered predictions through Google Maps, Autopilot features in planes, ridesharing applications etc. are day-to-day examples where the use of AI is done in an extensive manner.

There is no doubt about the fact that machine learning and deep learning has changes the face of the technology and with such moving and changing dynamics arises new issues and challenges. A draft report of the European Parliament to the Commission on Civil law Rules on Robotics stated that in future, there will be no stratum of society which will be untouched by AI. AI has been overlapping with the ambit of Intellectual Property Rights (IPR) and it raises many issues concerning the laws and methodology. 

IPR is an important tool to protect innovation and provide economic benefit for the intellectual work and AI has turned out to be a new dimension. WIPO recognizes three categories of AI- Expert Systems, which solve problems in specific field of knowledge such as medical conditions, Perception Systems, that allow the technology to perceive world with sense of hearing and sight and Natural language Systems, that requires a dictionary database to learn the meaning of the words.Trademark, Copyright and Patent laws exclusively need to be examined in light of the developing and rapid use of AI. 

Development of AI in IPR

According to John McCarthy, “AI is the science and building of making shrewd machines, particularly clever PC programs.” The concept of Artificial Intelligence ages back to over a century. The Greek myths of Antiquity first dealt with the concept of robots and artificial humans. There have been various important advancements in the field of AI ever since which include improvement and development of Turning Tests which assess the insight by Alan Turning and ELIZA which is a characteristic language used in preparing PCs. Looking back, the past 3 decades have been significant for AI.

The U.S. forces deployed DART in 1991 at the time of the Gulf War. It was an automated logistics planning and scheduling tool. In 2005, STANLEY, a self-driving car, won the DARPA Grand Challenge. In 2017, Sophia, a humanoid robot, became a citizen of Saudi Arabia. This raised the questions of AI machines having the same rights as that of humans.  A ruling from the San Francisco court in the case of Naruto v. Slater, popularly called “The Monkey Selfie Case”, denied copyright requests for a selfie taking macaque monkey and also represented the stand towards AI. This case gave rise to more questions than answers. 

The current sciences have information that is significant but do not amount to a creation according to the traditional definitions. This adds up to the fact that AI is booming around the world, along with the advancement in the economy, and the necessity to form and review its basic structure, which includes the ownership, licenses, abstract etc. is more important now than ever. 

To know more about IPR please visit 

Current situation

Currently, there is no particular law (in India as well as abroad) that answers the big question, “Who possesses the IP rights for a substance made by their invention?” The present law only regards humans as creators and hence, as IPR holders and infringers. This raises the issues of analyzing the status of use of these AIs in the future. It has also raised a lot of obligations with respect to the work created by AI’s.

United Kingdom

The IP rights of a creator i.e., the person who makes/ creates a substance are characterized under the Copyright Designs and Patents Act, 1988 (the CDPA). When an AI or a machine creates a substance, the IP rights under the CDPA are held by the creator of the machine, provided that the person is in direct control of the working of the machine with AI. 

In situations where the substance is produced by AI and there is no human creator, the IP rights will be held by the person making the game plans by which the work (AI) is produced. Hence, under the UK law, AI is not viewed as a creator. 

United States of America

The Copyright Act in USA protects original expression and not the ideas behind those expressions. In the case of protection of AI data, there is no particular provision that has been provided but the Act states that for a copyright requires ‘an original work of authorship’. The author has been interpreted through the US courts as a person or human being and thus while using AI, for any copyright protection, there should be human contribution in creative process. The AI being used is regarded as a tool for the process of copyright.   

China

In China AI is considered to include algorithms and data which are protected by intellectual property. Under Chinese law, the copyright of AI protects not only the expression of the algorithm but also the compilation of data. For the protection of the core idea of AI, patent is used. Further, such algorithms are to be protected by invention patents and not by the utility model patents. China has been able to put up a better structure for protection of AI. Though there are still ambiguities lying with respect to the ownership, especially in case of multiple stakeholders.  

Japan

Japan has been very advanced in regulating the AI with IPR. It had formulated the ‘AI Strategy 2019 AI for Everyone- People, Industries, Regions and Governments (2019)’ which focused on identification of problems and examination in future. It also made changes in its Copyright Act by amending Article 30-4, 47-4 and 47-5 which introduced the flexible limitation provisions for technologies like IoT and AI. Furthermore, the database is also protected through Article 12-2 (1) of the Copyright Act and it states that even if the whole database is protected, the rights of author of a work that forms a part of that database will not be affected. However, there are still ambiguities with respect to ownership of data and the issues relating to patent still needs to be addressed.  

Way Ahead 

The next generation of AI is expected to deal with more practical situations, where there would be no access to any data. Instead, intelligent agents must self-learn through trial and error to make decisions bearing in mind the long-term payoffs. Thus, the next-generation AI, which is still not fully realized in practice, would have more autonomy and sophistication in decision-making. The scope of AI is not complete without robotics and autonomous systems, characterized by the physical embodiment of intelligence in the real world. One may view embodiment as an independent facet of AI. Nevertheless, it is the emphasis on embodiment that closes the loop with the real world through sensors and actuators, which helps AI to be in control.

The significance of AI in future is so strong that there is a desperate need to regulate it before there is chaos. The inclusion of AI in almost every sector of our lives comes with its own risks and liabilities. Since the data holding capacity of these AIs is so large in number, it can be manipulated and used in certain wrong ways as well. Thus, the ownership and liability of a person needs to be decided. IPR is such a gateway which describes the ownership of such work and the originality by a person. In case of AI, it won’t be useful just for the owner to gain from the profits of such usage, but it will also be easy to maintain and establish the liability in case the results of AI are destructive or cause serious harm.

The future holds much more challenging aspects in case of IPR. Even the developed countries have not been able to turn around the whole scenario of AI and IPR. However, the courts in these countries have taken cognizance of the facts and have given judgments on the issues. The issues of ownership and economic profits in invention of AI while maintaining the market balance and encouraging new inventions will persist.            

AI & IPR in India

The development model in India includes a huge amount of technological advancement which includes AI as well. The use of AI is not limited to social media or entertainment but has accelerated to retail as well. From online shopping to the use of online car services, the country saw a rapid change in the technology. The issues in a developing country like India are of much more concern as it is the basic infrastructure which needs to be revised upon. There are well-established patent and copyright laws in India.

However, there is no particular act or provision to regulate AI specifically. The existing laws do not cover the ambit of AI and are based on the old intellectual property types like books, creative writing and discoveries. The ambit of AI is much more complex and needs to be addressed in a particular way, different from the existing regime. Under the Patents Act of 1970, computer programs, business methods or mathematical formulae are not considered as patentable inventions.

Furthermore, the terms ‘patentee’ under Section 2 (p) of the said Act and ‘person interested’ under section 2 (t) of the said Act creates a barrier to include AI in its scope. The Act specifically terms out the patentee of any other person interested to be human.  

Under the Copyright Act, there are two basic doctrines which define the originality of the work under this Act- Sweat of the Brow Doctrine and Modicum of Creativity. Since the doctrine states that a minimum degree of creativity is also acceptable, the original work of AI can be included in it. However, the rights of copyright are given to the ‘author’ of the work done under section 2 (d) of the Act. Author in this act has been implied to be a human or legal person, thus, making the idea of machine to be protected under this act restricted.

The current regime and laws are not in consonance with the upcoming and even existing dynamics of technology. In a country with second largest population and majority of the people using social media sites and online shopping, it is crucial that the laws should be amended according to the new structure. The new technologies include various features like Amazon’s AI product ‘Alexa’ is being used as a security measure to lock doors at homes. In case of failure due to confusion or misunderstanding of AI, there are various questions such as- who will be made liable? Can the liability be shifted to the user? Furthermore, any new invention based on same algorithm or same concept may hamper with the rights of the original owner. This becomes a serious issue. On one hand, it can discourage the start-ups to be inventive and thus, ruining the whole purpose and existence of IPR at the first place and on the other hand it can cause a series of litigation and chaos in the IPR sector.    

Copyright and AI

Copyright is possessed by the creator of literary/artistic or musical work which allows the creator the exclusive rights to the sale/ use/ distribution of the work. 

To know more about Copyright please visit

In the case of Burrow Gilles Lithographic Co. v. Sarony, the issue was whether a copyright protection can be granted to a photograph as there was a conflict between creative and mechanical labor. The court held that copyright protection can be granted to a product which was the output of a machine, which in turn, is the output of human creation. It also narrowed their scope of protection by stating that mechanical labor cannot be held to be creative. Therefore, if a strict approach like this were to be applied to AI systems, granting copyright for works created by them, would be difficult.

In another similar judgment of Bleistein v. Donaldson Lithographing Co, the court differentiated between the works of a human from that of something artificial. Justice Holmes stressed on human nature as an important aspect to create a copyrightable work.  The Court emphasized on the matter that there was no scope for anything that was not a product of man’s creativity to be copyrighted. 

Even if countries agreed to granting copyrights to the works of an AI, the question of who gets that copyright remains cryptic and abstract. This is because the current status of law requires a legal personhood of a right holder, something which an AI lacks, unless its creator is granted that on its behalf. 

EU & United States

Firstly, there has to be a clear distinction between AI created works and AI assisted works.  In the former case, there is no protection because there is an absence of human intellect and in the latter case, the copyright holder is the human who created the substance. Additionally, the requirement of the AI’s work being original is questionable. Works entirely made by the code of a computer program are not protected and they are free to use by anyone.

UK

Contrary to the practice followed in the EU and USA, the computer-generated works (CGW) regime is implemented in the UK. According to this regime, the programmer of the AI software is the copyright holder of the machine’s output.

Patents and AI

In today’s times, AI enabled systems are creating and inventing new outputs, which is beyond the knowledge and understanding of its inventor itself, and this is considered a huge development but it also leads to ambiguities from the perspective of patent law. A patent is the exclusive right granted to the inventor for his invention (product/process) if the invention is novel, non-obvious and useful. Granting of a patent to the patent holder, excludes others from making, selling, distributing the invention without the grant of licenses from the patent holder. 

AI is booming to an extent that AI invented machines are producing results which could qualify as patentable inventions.

EU & United States

Under the US law, an ‘inventor’ is defined as an individual or a set of individuals who invented or discovered the subject matter of the invention. This definition eliminates any intention of the US to include inventions made by AI to be patented. 

However, the European Union is working towards expanding their national laws to include the copyrightable works produced by computers and other devices/ machinery under the category of ‘own intellectual creation’ but no efforts are being made with respect to patents by AI systems and robots in this field.  

UK

The UK law speaks of no provisions relating to patenting of works created by an AI software. In a case where a patent application was filed in UKIPO and DABUS (patented AI system) created by Dr. Stephen Thaler, (also called the ‘creativity Machine) was called the inventor. This application was declined by the UKIPO stating that DABUS was a machine and not a natural person and hence, the inventions created by DABUS will not be considered. The UK provisions provide that the inventor must be a human person and the Hearing Officer of the case held that there was no law that allows the transfer of ownership of the invention from the AI inventor to the owner of the AI. He also held that the inventor of the AI cannot take the credit for the invention made by the AI in the above case. Hence, the ambiguities regarding this case still persist.

Trademark & AI

Trademark law strives to eliminate any confusion with respect to logo, appearance or packaging or any other marks which identifies a particular brand or company so that there is no confusion among the consumers. It is difficult to analyse how AI can infringe trademark and issues like in case of patent and copyright can arise. There have been instances where the question of infringement of trademark with respect to AI arose. In the case of Louis Vuitton v. Google France, there was an issue of keyword advertising and the automated choices which are made through Google and it was alleged to be infringing the trademark of the petitioner.

However, the court held that there is no infringement unless the party itself took an active part in it.  The case of Lush v Amazon brought out the need to structure the trademark laws as well as for future issues. Lush had not allowed Amazon to sell its products on the website. Through a bidding process, Amazon had bought the keyword ‘Lush’. Therefore, even if Lush was searched through Google, it would show Amazon advertisements as well. And though there were no sales on the website, the AI would keep showing the similar products based on the searched keyword. A suit for infringement of trademark was filed by Lush and the court found Amazon guilty for the same.

With the increasing use of AI in retail and business models as well as security and payment methods, this situation can get much worse. The case can deteriorate more if there are AI based consumers. AI is based on algorithms and uses the data by identification of prior choice. If an AI becomes a consumer there is a huge chance of confusion in respect with trademark and this situation can lead to a rigorous amount of litigation. 

There are no particular laws in any country with this respect and it is the need of the hour that certain regulations are made to avoid any such confusion in coming years.   

If AI owns IP

In a situation where the AI is the IP right holder of an invention or a creation, questions arise relating to the aspect of infringement.

Firstly, if AI is given the same position as an individual for creating or inventing a work, then it should also be made to enter the field of infringement and enforcement. An AI software should be sued for infringement and should be able to enter into legal contracts on its own which does not seem possible moreover, appropriate. This proves the fact that AI cannot be a legal entity. 

Secondly, the question of accountability arises when an AI infringes the right of a 3rd party. Especially in the cases of copyright, if an AI copies the work of an author, given that it is easier as all the works are available on the internet, the hurdle of showing that the infringer had access to the protected work might be much easier to overcome. 

Thirdly, there is an issue of transparency of AI systems with respect to holding of IP rights by them. If a situation arises where AI systems are protected under trade secrets, this could be an obstacle to the transparency of the AI systems. As times are progressing, there is more need and importance of transparency and responsibility for the process of decision-making.

Hence, the question of the hour is, “how to deal with transparency in cases where a machine learning process involves multiple data sources, dynamic development, and elements that are opaque, whether for technological or legal reasons?”

Conclusion

Artificial Intelligence has become an integral part of human lives. With the use of AI, data can be modified or collected in a much better and time-efficient manner. The usage has boomed with the use of new technological tools. Therefore, it becomes an urgent need to make proper laws concerning the same. AI is a technology which is moving at a fast pace and it is crucial to examine and analyse the issues and challenges which might surface with it. IPR is already an expanding area. With technological approach there are many factors which are being introduced under its ambit.

AI being one of them should also be looked into. The present situation of AI and IPR is challenging. The implementation of IPR in AI with procedure and safety logs at hand is a real problem. There is also an issue of understanding the different features of AI, which becomes a more rigid problem in developing countries like India, which are still going through rigorous changes in technology.

Presently, it is through the interpretation of the courts that the issues revolving around AI and IPR are being addressed. But there is a need for structured, analysed and clear rules and regulations. Amendments should be made in existing IPR laws to address the issue of AI as well. With the use of AI, there can be more benefit for future inventions. India has been potentially looking forward for ways to do. Niti Aayog in a discussion paper in 2018, described about the importance of AI in healthcare, education, infrastructure etc.

The benefit of such technology during pandemic was also recognized and it was suggested that AI should be made to meet the criterion under the Indian Patent Act to facilitate faster operation in data processing, screening, publication, examination, medical support, hearings and application filing during pandemic. However, there is still a long road to follow. Not only the laws need to be regulated but better infrastructure for the implication of such rules also needs to be determined.

Suggestions

There is a desperate need, now more than ever, to formulate IP laws which can secure the developments of AI innovation and reward new works and inventions through the copyright or patent award. 

• A specific test has to be formulated which can differentiate between AI created works and AI- aided works. The accurate IP holder can be determined, thereon. 
• The patent law clearly demarcates between an inventor and an invention but the category under which AI systems fall is not yet determined. The law has to be clearer and more specific and include such provisions in an understandable language. 
• Similarly, the definition of authorship under the copyright act should be examined and changes according to the changing dynamics. 
• There have been ambiguities in trademark laws as well. The status of AI as a customer and the functioning of AI, especially in case where human common sense plays a major role need to be defined. 
• WIPO has already taken cognizance of the upcoming issues with AI and the same has been discussed through various means, however, proper policy should be formulated in an international level.  
• A specific act has to be passed which deals with data protection with respect to the AI software. It must cover all the civil and criminal liabilities and offences which amount to the same. 
• IP sharing between the inventor of the AI and the AI itself can be a possibility in the years to come. It will become an essential part of the general advancement plan and maintainability. The future is loaded with indefinite possibilities and interesting motoring. 

LawSikho has created a telegram group for exchanging legal knowledge, referrals and various opportunities. You can click on this link and join: