The rise in AI based patent applications and grants in the field of telecommunication and medical sciences is more than ever before. Although the issue of authorship or ownership by AI has been dealt several times in the past. Like an AI system, DABUS filed patent applications as an inventor in UK, USA and Europe, all of which were rejected on the grounds that it was not a legal person as required by most IPR regimes and later, a similar application was also filed by Google’s digital news initiative foreign AI system which creates local news content with the help of human made templates, but no conclusion has been drawn so far and this of ownership issue is up for debates in many international platforms.
It is important to note that the main motive behind IP rights is to secure the rights of the owners and encourage them by incentivizing their work, theory which does not apply on AI as it does not need any motivation or reward to work. However, it may propel humans to create such technologies if given the due credit!
AI and Copyright
There are ever increasing examples of AI generated artistic works such as The Next Rem brandt portrait in Netherlands, a short novel written by a Japanese computer program in 2016 reached the second round of a national literary prize. An artificial Intelligence company, Deep Mind, owned by Google created a software that can listen to recordings and generate music.
Different approaches are adopted by different jurisdictions. Such as the United States Copyright Office updated its interpretation of “authorship” in 2016 to clarify that it will not register works produced by a machine or a mere mechanical process that operates randomly or automatically. Similarly, in a recent Australian case, a court declared that a work generated with the intervention of a computer could not be protected by copyright because it was not produced by a human. Thus, denying copyrights to AI.
However, countries like India, UK, Ireland, Hong Kong and New Zealand grant the authorship to the programmer ensuring that companies keep investing in the technology, ensured that they will get a return on their investment.
AI and Patent
The growing numbers of AI inventors such as Philyra – an AI product composition system using machine learning algorithms for crafting fragrances poses a need for patentability of AI generated inventions.
Though under U.S Patent Law, an ‘inventor’ is defined as an individual or a set of individuals who invent or discover the subject matter of the invention. Eliminating the possibility of granting patents to anything non-human, the patent procedures are evolving giving rise to numerous AI patent applications with USA and China being the leading countries in this domain. Filings under WIPO’s Patent Cooperation Treaty (PCT System) represent 20% of the total number of AI-related filings. It is evident from the statistics that AI is the most prominent in the patents field as opposed to any other IP regime.
AI and Trademark
The marketing world dealing with product branding influencing consumer decisions especially in e-commerce companies rely largely on AI for recommending products to the buyers proves the high dependability of brand value appraisal on AI. Therefore, it involves a high amount of risk in giving trademarks rights at its disposal. The impact of AI on TM laws is also being addressed at a global level as discussed in an article in WIPO Magazine, 2020.
Challenges for giving IP rights to AI
WIPO launched the first of its kind, global initiative in the field of extending IPR to AI inviting proposals and contributions from member states. The aforementioned WIPO Discussion Paper, December 2019 brought up and later dealt with the issues regarding the matter at hand and I have made a humble attempt to suggest my opinion on the questions raised as follows:-
- How to deal with the protection of patent applications generated autonomously by AI and should it become available to public?
- In my opinion the AI supported and assisted works involving human intervention should be incentivized to the extent of human contribution in the artwork. Whereas the solely AI- generated work should be brought under public domain
- How to meet the disclosure requirement and whether the initial disclosure would suffice the continuously changing algorithms in machine learning?
- Patents should not be granted to AI assisted or supported works wherein the transparency of algorithms cannot be maintained for public access to innovate further. The description of algorithms and the combination of training data needs to be disclosed and updated even for continuously evolving machine learning.
- How to determine the authorship of autonomously generated literary and artistic works by AI for copyright?
- The AI generated work needs to be categorized into three broad categories, that are AI-supported, AI-assisted, and AI-generated in terms of the degree of human intervention and then decide the authorship for autonomously generated artistic works
- Does the existing IPR regime suffice to accommodate the complexities of identifying and protecting AI generated data?
- The current legislative framework can accommodate the patentability of AI assisted inventions, whereas to deal with the AI-generated inventions we need universal harmonization for adopting an internationally agreed standard
- Who should be held liable in terms of IP administration?
- The owner of the AI developer should be held vicariously liable for all the criminal violations of IP rights committed by their creation
The issue of requirement of incentivization can be solved by utilising the funds generated through commercialisation of patent, copyright and/or trademark protected AI generated work. The funds thus acquired infested in enhancing the AI itself. This resolution would suitably incentivize the human creator of the AI and would also promote technological advancements in the country and the world at large paving the way for an advanced yet systematically protected digital age. What can be better than compensating the AI for its efficient work by investing the remuneration generated through it on the AI itself?
On the other hand, it is also imperative to note that AI inventors develop at much higher and uncontrolled pace through machine learning as compared to human inventors. Thus, providing patents to inventors would create social costs and monopolies stifling the entry of new human ventures and thereby hampering innovation.
There are many challenges arising at the intersection of IPR and AI based on disclosure of work format, identification of infringement and paternity of work. We are left with no other choice than to mend our IP laws to promote an inclusive environment for the interface of IPR and AI
The existing conventional IP laws have a lot of room to accommodate these challenges and therefore need amendments for the same. WIPO has taken the great initiative of addressing the issues arising due to overlapping of IP laws with complex nature of AI based inventions and it needs to be taken further actively by member states, concluding to the terms suiting both human creators of AI and the machines itself while shielding them from criminal liabilities at reasonable grounds.
Granting IP rights to AI is a novel approach and is not incorporated in the IPR laws of most jurisdictions wherein the AI systems are treated merely as software incapable of holding accountability or ownership of independently created work.