The World Intellectual Property Organisation (WIPO) in 2019 published a report discussing trends in patent applications and grants, noting that of late, there has been a rise in AI-related applications in the fields of telecommunications, transportation, and life and medical sciences. These patents have involved technology relating to natural language processing, speech processing and computer vision. In 2017, an AI system named DABUS was named as the inventor in the patent applications filed in United Kingdom, USA and Europe. However, the same was rejected in all three jurisdictions on the account of it not being a legal person as required by most Intellectual Property Rights (IPR) regimes. Similarly, Google’s Digital News Initiative has funded the creation by the UK’s Press Association and Urbs Media of Reporters and Data and Robots (RADAR), an AI system which will help create local news content based on templates created by real journalists across various genres. This surge in patent filings claiming authorship or ownership by AI over a period of years ,have raised interesting questions and discourse in the world of IPRs, as to whether AI can be treated as the author or creator of innovations.

A separate issue that is considered in the context of AI and IPRs is the practical difficulty of the patentability of AI systems in the context of the subject-matter eligibility standard. In most jurisdictions, algorithms by themselves qualify as a vague system lacking technical character, and hence cannot be protected under the IP laws, unless it has been given a technical character in the form of effective software. However, in response to the changes in technology and rising applications, IP regulators in different countries have also come up with guidelines regarding examination, and initiatives to encourage patent protection in these areas of technology. They provide clarity over the eligibility of algorithms that can come under the purview of patentability, for them to then be tested on merits such as novelty and enablement.

Another important debate is on issues related to ownership/user/authorship rights of content, and inventions that are autonomously generated by AI systems. At the present stage of development, though, examples of content generated by absolutely autonomous AI systems are few and far between; we are still quite a way from ‘independent acting’ computers being more ubiquitous in society. Nevertheless, this gives rise to a debate about the patentability/copyrightability of the inventions and content that is created using AI systems as a tool. The current IPR laws accord rights to entities accorded a legal personhood (whether natural or corporate); it is for this reason that in most jurisdictions, at present, an AI system is precluded from the grant of such protection. However, since AI technology is still very nascent, and there is ambiguity around the definitions of AI and “autonomy”, there appears to be some difficulty around legislating on this point.

This further makes it difficult to ascertain the standards associated with IPR laws such as duration of protection, identification of beneficiary for licencing remuneration, differentiating between human and AI creations etc. While there are many challenges, one argument for allowing computers to be classified as inventors/authors and afforded with IP protection is the ‘incentive theory’. While this may not be any motivation for computers themselves, it will continue to incentivize humans to produce such technologies as they understand the benefits emerging due to IPR protection.

In September 2019, WIPO organised a conference to discuss the impact of AI on IP policy of various countries and the relevant questions to set the foundation for better informed policymaking by member states. Pursuant to the conference, WIPO published a Draft Discussion Paper on IP and AI in December 2019 (WIPO Discussion Paper) inviting member states and other interested parties to provide comments and suggestions. The WIPO Discussion Paper identifies thirteen issues that relate to the issue of AI and IP policy:

• Issue 1 pertains to ownership and inventorship. It deals with issues such as whether the law should permit or require that the AI application be named as the inventor or whether this should necessarily be a human. Further, it considers the practical challenges of whether there should be any indicators of which human ownership or authorship should be attributed to, if AI systems cannot be given ownership, i.e. whether this decision should be left to private arrangements, such as corporate policy, with the possibility of judicial review by appeal in accordance with existing laws concerning disputes over inventorship. Finally under issue 1 the WIPO Discussion Paper asks comments of the member states on the question –‘Should the law exclude from the availability of patent protection any invention that has been generated autonomously by an AI application?’.
• Issue 2 of the WIPO Discussion Paper is about patentable subject matter and patentability guidelines. Here, it considers the issues of whether inventions autonomously generated by an AI application ought to be excluded from IPR laws, whether specific provisions should be introduced for inventions assisted by AI (or if they should be treated in the same way as other computer-assisted inventions), whether patent examination guidelines need to be amended for AI assisted inventions, etc. Under issue 3 the WIPO Discussion Paper explores the issue of understanding the inventive step test that needs to be met for the invention to be granted a patent in the context of AI inventions.
• Issue 4 deals with disclosure of the technology, and whether AI-assisted or AI-generated inventions present any challenges in the disclosure requirement; further, it considers whether the initial disclosure requirement would be sufficient where the algorithm continually changes over time through machine learning; how to treat data used to train an algorithm; and whether human expertise used to select data and to train the algorithm be required to be disclosed.
• Issue 5 relates to general policy considerations such as whether a sui generis IPR system should be considered for AI generated inventions, and whether the interface between AI and IPRs should be considered at a later stage once AI technology itself is more advanced or better understood.
• Issue 6 relates to copyright and discusses authorship and ownership issues, such as whether copyright be attributed to original literary and artistic works that are autonomously generated by AI; in whom should copyright in an AI-generated work vest; whether the issue of according legal personality to an AI application should be considered, where it creates original works autonomously; and whether a separate sui generis system of protection ought to be envisaged for original literary and artistic works autonomously generated by AI.
• Issue 7 pertains to IPR infringements and seeks to understand whether the use of the data subsisting in copyright works without authorization for machine learning would constitute an infringement of copyright and what impact that would have on the development of AI and on the free flow of data to improve innovation in AI; whether an exception should be made for limited types of use of such data in machine learning, such as the use in non-commercial user-generated works or the use for research; how would existing exceptions for text and data mining interact with such infringement; whether a licensing system would be useful as an alternate to copyright infringement and whether the unauthorized use of data subsisting in copyright works for machine learning can be detected and enforced, in particular when a large number of copyright works are created by AI. 
• Issue 8 considers the issue of ‘deep fakes’ or ‘the generation of simulated likenesses of persons and their attributes, such as voice and appearance’, and whether copyright can subsist in deep fakes themselves; whether there should be a system of equitable remuneration for persons whose likenesses and “performances” are used in a deep fake;
• Issue 9 relates to whether there are seen or unforeseen consequences of copyright on bias in AI applications; whether the dignity of human creation should be prized as a right over and above innovation in AI;
• Issue 10 considers whether there should be a new set of IPRs in data or whether the existing regime of IPR laws are sufficient; what kinds of data would be protected under such new rights, if created; whether certain qualities in the data such as commercial value, or protection against certain kinds of activities should be the defining characteristic for these new rights; how would such rights interact with existing rights and how would they be enforced.
• Issue 11 considers industrial designs and looks into questions such as whether design protection should be accorded to an original design that has been produced autonomously by an AI application, or whether a human designer is required;
• Issue 12 addresses capacity building, to address the containment or the reduction in the technology gap in AI capacity and whether any policy measures are required in this regard.
• Issue 13 of the WIPO Discussion Paper pertains to accountability for the decisions in IP administration.     
• In light of the above, the current chapter maps IPR policy vis-à-vis AI of various countries, and their views on these issues in their respective jurisdictions. In most cases, the legislative framework treats AI systems as equivalent to software, and therefore offers limited protection; further, in most jurisdictions, the issue of granting AI itself authorship status is a novel one and is not something that is accounted for in the existing IPR system.