WhatsApp Image 2021-07-18 at 11.05.59 AM (1)



The machine learning systems today are mere robots providing automatic assistance to humans but with the advancement of technology, the role of AIs is bound to expand manifold. The current scenario depicts a restrictive role of AI which recommends products according to analyses made by the programme factoring in the buyers’ preferences, cash in hand, search history, prior purchases. The chatbots used as shopping assistants by various brands like H&M, have expanded the role as they generate outputs based on personal information feeds. New interfaces of various companies with more personalised outputs based on skin colour, hair complexion, and physical attributes would play more than a suggestive role as the consumers will rely and trust on the choices of the system (logical parameters computing viable choices of the consumer). This opens the cracks for liability issues in cases of infringed products. There have been various cases reported on Snapdeal where the consumers’ were suggested branded products based on the parameters used by the platform, which turned out to be first-copies of original brands like Nike, Adidas etc.  In the wake of recent reports about fake products being delivered to consumers, Amazon launched its Project-Zero anti-counterfeiting pilot which uses AI. The online marketplaces may see a surge in the usage of AI systems requiring filters for discriminating brands and fake products which may otherwise lead to trademark enforcement issues. The concept of average consumer his/her awareness and judgement in AI software compared to humans boils down to the argument in the case of online marketplaces as there is no physical attribute present to compare original with fake. Thus the decision of product choice is totally dependent on AI system, considering the example of a consumer willing to buy a PUMA shoes and opting for the same but it may so happen that due to indiscrimination the AI may deliver the product from a fake retailer due to phonetic and visual similarity (the major inputs used by an AI used while in selecting a product).

Ever since e-commerce websites and applications have employed AI systems in their operations, trademark enforcement agencies have been going haywire about its liability in the current as well as the future regimes. As the role of AI’s expands, from suggestive to recommendatory and furthermore in the years to come, so does the level of complexities in law. The answer, as they say, is provided in the question itself. The liability of a party in any transaction is evaluated on the basis of its involvement in the same. In the present scenario the AI system provides the multiple choices available to a consumer. The decision to choose however still remains with the person making purchases not with the machine and therefore the liability still remains on platform and retailers. For example a counterfeit product from Shoppers Stop or ZARA doesn’t hold the salesman liable as he’s an innocent infringer suggesting products in good faith. Therefore, an AI no matter how much advanced it is would never take the decision of buying a product; the right exclusively remains with the purchaser. Mere display of a counterfeit products can lead to abetment is sale of a fake product which still shifts the liability on the platform. Thus the platform and not the AI used can be held liable in any case.


The sweeping changes in the Artificial Intelligence arena in the self learning, cloud computing paradigm like quantum supremacy have raised various pertinent questions with respect to the intellectual property rights affecting there from not only in the trademark sector but also in various other IP like copyright and patents. The major question that is posed by the self learning tools is with respect to the liability of A.I. in cases of infringement. In the case of trademarks the first question anyone looks at is deceptive and visual similarity that would be created when two similar brands, one counterfeit and other original are posed against each other and in the same case if the A.I. mistakenly orders the counterfeit product. The AIs today have the capacity of working and functioning with human aid but in the future the element of human involvement would reduce and the action of self learning machines will become more autonomous.

The more autonomy with machines means it is difficult for the person to ascertain the liability in cases of infringement. Machines like Dash and other replenishment services provide machines which function on the data collected by them by earlier purchases made or the stock levels provided by the consumer. Hence in such cases the chances of ordering counterfeit products increases. There are basically two opinions regarding the liability to be fixed in cases of infringement the policy makers who believe machine as a form of platform or intermediary believe that it should be placed under the safe harbour provision mentioned under section 79 of the IT Act which states intermediaries are supposed to be exempted if anything is posted or uploaded on the platform, the transmission is not initiated, is not selected receiver of transmission or modify the information by the host itself. Thus in the present scenario If we go by the natural law doctrine the AI or any other machine cannot be held responsible in the court of law until and unless it does not have any constitutional rights like in Saudi Arabia, but even in those cases since it’s a machine there cannot be any harm inflicted on the machine. The bottom line being that ultimately a machine does not hold good in the court of law. The civil law and criminal law remedies with respect to AI are still not adopted in the domestic laws of the country. The statutes have to incorporate those remedies and a specific structure of remedies has to be framed in order to bring all the other laws in synchronisation.

Further the other opinion states that the programmer or the company which holds the ownership of the machine should be held liable in the cases of infringement by the AI. The counterfeit products which would be served by the faulty machines or the programmes mentioned therein. This is going to shift the burden on too many people and it would be really difficult to ascertain the real liability, though joint-liability can be placed. But still there is a problem of ascertainment of liability with respect to 



Share on facebook
Share on twitter
Share on pinterest

Leave a Reply

Your email address will not be published. Required fields are marked *

 IPTSE Webinar


IP in Career Development

7th July 2020 – 1 to 2 pm

Introduction to IPRs

14th July 2020 – 1 to 2 pm

Copyright Protection – Part 1

21st July 2020 – 1 to 2 pm

Copyright protection – Part 2 (Case Laws)

28st July 2020 – 1 to 2 pm

Design Protection – Part 1 

4th August 2020 – 1 to 2 pm

Design Protection – Part 2
(Case Laws)

11th August 2020 – 1 to 2 pm

Trademark Protection – Part 1

18th August 2020 – 1 to 2 pm

Trademark Protection – Part 2
(Case laws)

28th August 2020 – 1 to 2 pm

Trade secret
8th September 2020 – 1 to 2 pm
Design Protection – Part 1 
15th September 2020 – 1 to 2 pm


Webinar on Role of Intellectual Property In Our Daily Lives
Play Video
Play Video

SMEs and Startups Webinar


Trade Mark & Copyright Protection for MSMEs

1st July 2020 – 1 to 2 pm

Effective utilization of the IP system for MSME

10th July 2020 – 1 to 2 pm

IP value extraction and commercialization

19th July 2020 – 1 to 2 pm


Play Video
Play Video
Play Video