This separate legislation, Digital Competition Act, prescribes a certain cut-off for the law to be applicable and marks certain practices such as bundling, tying, self-preferencing and cross-utilization of data across platforms as anti-competitive in the market.
Background: The advent of Generative AI has taken the development of technology to the next level. As per one estimate, as high as 40% of the jobs are likely to be impacted by the advent of AI. The ability of AI chatbots to process terabytes of data within a couple of seconds and answer even subjective questions with almost pin-point accuracy has surprised many. It is now known that Gen AI machines continually assign ‘weights’ to the existing knowledge (logic), which makes this processing possible at a lightning speed even at a mass level.
For instance, an AI bot would be able to solve mathematical equations much quicker due to the higher ‘weight’ assigned to the underlying logic of arithmetic calculation, as compared to a problem rooted in cultural context. The broader understanding is also that since the AI industry pivots around data, subsequent competition concerns that arise would also fall in line with conventional digital markets. In fact, as per Autoridade da Concorrência (Portugal), Gen AI is the poster child of all competition concerns in digital markets.
Analysis: One factor that sets the AI industry apart is its nascent nature. The conventional understanding of economic regulation is that markets have to develop first for regulations to ply. This understanding is, however, now under challenge; the Standing Committee (Finance) in one of its reports has held that tech markets tip-off within 3-5 years. This means authorities must be on a vigil so that timely intervention is possible.
This vigilance may be necessary in the context of AI and the fast evolution of the supply chain in these markets. As per a paper published by the Centre, it is suggested that the number of servers in the AI value chain is highly customizable, with the exact number being determined based on the nature of the task involved. In other words, the supply chain involved would be more complex in the case of medical systems as compared to logistical operations given the underlying task and sensibility.
Another key factor in AI is the computation power available with companies. This ability gains prominence as one of the apprehensions cast by various antitrust authorities is that this component is likely to be dominated by incumbents like Google, Microsoft and Amazon, leaving smaller companies at their mercy when it comes to settling terms and conditions of any agreement.
There is already a shortage being reported in the accelerator-chip market in favour of Nvidia. Evidence has already started emerging that these Big Tech companies are signing exclusive agreements which would make them exclusive providers of cloud services to tech companies. This trend is all-the-more worrying in India as even though there are more than 60 Gen AI startups, none of them is able to build its own foundation model.
Another concern that warrants attention is the discourse on open-source technology. It is typically understood as a dataset developed by a company that could be used by anyone on the internet, thereby enhancing the public good. The devil, however, lies in the detail. A key concern with open-source technology is that it is mostly applicable to end-consumers, thereby keeping its business usage at bay.
This results in the creation of entry barriers for businesses working in the AI supply chain. Another issue is around the proprietary rights held by the creator of such technologies, which puts into question the very definition of anything called open source. It wouldn’t be out of context to suggest that this terminology is susceptible to regulatory abuse as it gives a false impression of fair play when it doesn’t exist.
Last but not the least, technical expertise in the form of human programmers lies at the center of AI development. It is now known that this technology could very well be used to achieve destructive ends if not monitored properly. The Indian government has issued an advisory to that extent. It is, thus, imperative that AI companies, apart from ensuring development, must also make sure that chatbot training is supervised by human raters to curtail adverse effects.
Conclusion: AI has made markets more efficient. It may be an understatement to suggest that a major potential present within this industry is yet to be unlocked. A level playing field and equal opportunity to prosper, however, remains a policy imperative. This objective is even enshrined in the Sustainable Development Goals. The need of the hour may be for competition authorities to come out of the ‘ex-ante vs ex-post’ frame and closely follow the developments in the market.
The Competition Commission has already formed a Digital Markets Unit to that end. This would allow them to intervene timely in case of any wrongdoings and ensure that efficient markets remain sacrosanct. This approach will not only serve national interest but also ensure that development of technology is given the due impetus it deserves at the global stage to serve the common good.
The authors are, respectively, founding director at the Centre for Competition Law and Economics (CCLE), and dean at SRM University, Sonepat (Haryana).
You are on Mint! India’s #1 news destination (Source: Press Gazette). To learn more about our business coverage and market insights Click Here!
Download The Mint News App to get Daily Market Updates.
Published: 24 May 2024, 02:00 PM IST
#Promoting #competition #digital #markets #challenge