Light-touch approach apt for AI regulation

The DCB seeks to regulate on an ex-ante basis a cohort of 10 core digital services (CDS) such as online search engines (e.g. Google Search), online social networking (Facebook), video-sharing platforms (YouTube), operating systems (iOS), etc. Of particular interest here is that the MDI is required to examine aspects like the current dynamics of these CDS, as well as the need to expand or reduce the CDS list, and the implications of the Bill on these services. The MDI is expected to consult various stakeholders including MSMEs and startups as well as large tech companies. Its recommendations will be of much significance in influencing further thinking; thus, its role is critical.

Although it hasn’t been specifically mentioned in the terms of reference for the study, an issue that has come up for discussion is whether AI is a fit candidate for inclusion in the list of CDS. What are the pros and cons of this debate?

AI is a young phenomenon. It has emerged like a storm and companies and governments alike are devoting much attention and investing billions of dollars into it. Its growth is phenomenal and almost every week there is a show-stopping development. In India alone, the AI market is growing rapidly; the Competition Commission of India’s (CCI) MDI study had expected the Indian AI market to grow from $11.17 billion in 2025 to $131 billion in seven years — at over 42% annually.

A Nasscom study has pointed to the exponential rise in the number of generative AI startups — growing almost four times in one year — and that AI is rapidly scaling in size, deepening in capability, and becoming strategically important to enterprises, governments, and the broader technology landscape. According to the Stanford Global AI Vibrancy Index, India ranks third behind only the US and China. New entrants are showing up both in AI applications and foundation models. This demonstrates not only the highly dynamic nature of the technology and its markets but also the low entry barriers for new players.

Indian consumers and startups have been significant beneficiaries of the AI spread in the country. Startups have relied heavily on ease of access to cloud services, open-source models, data availability, and also the integrated AI services from large firms. Early on, regulation could disrupt this rapidly evolving ecosystem with its cascading effect on innovation and the startup community.

The government has positioned itself as an active growth agent. Its focus has been on growth and innovation including through the IndiaAI mission. It has been trying to incentivise expansion in the fields of data (e.g India AIKosh), foundation models like Sarvam and BharatGen, and compute power like making high-end graphics processing units affordable.

Despite the rapid growth of AI, India is still way behind in aspects like computing power, local foundation models to match prominent global models emerging from the US (ChatGPT)and even China (DeepSeek and Kimi K3), large domestic data, and so on. At this nascent stage, the government has rightly shown little appetite to regulate the sector, preferring a light-touch, pro-growth, pro-innovation approach. Premature regulatory intervention can chill innovation, growth, and even investment which can choose less regulated countries. This effect is even more likely if the regulation is ex ante under the DCB instead of being ex post.

It is notable that even the CCI’s own study on AI markets (also undertaken by the MDI) never recommended ex-ante regulation. Instead, it suggested softer measures like self-audit by companies, greater transparency, and minimising entry barriers through the development of capacities such as AI computing power, building public data repositories, and AI skills. It also recommended building the CCI’s capacity and expertise in the field, which is admittedly a necessity both in numbers and in depth. The ministry of electronics and information technology has never favoured intrusive regulation of AI, concentrating instead on accelerating capacity and growth through its policy and structured incentive projects.

Other jurisdictions, including more mature authorities, have held back on ex ante regulation of AI. The European Commission (EC) was one of the first to introduce ex ante regulation of digital markets through the Digital Markets Act (DMA), which came into force three years ago. There have been contradictory views about the positive and harmful effects of such ex-ante regulation. The EC did its first review of the DMA recently and pointedly considered whether AI should be covered under the Act’s Core Platform Services list. The EC concluded that the current list of CPS should not be expanded at this time to include AI services.

Similarly, in Japan, the Mobile Software Competition Act does not identify AI as a stand-alone digital service to be brought under regulation. In the UK, the Digital Markets, Competition and Consumers Act does not include a specified list of digital services. Nonetheless, in relation to Google Gemini Assistant, the UK authority ruled after investigation that it should not be included within the scope of Google’s search services at this point.

These examples reflect that AI is not a separate product but rather a general-purpose technology that is getting embedded in existing products and services for enhancing their functions and user experience.

AI is now integrated into operating systems, workflows, physical products, and is even executing complex tasks; agentic AI tools are increasingly popular in this respect. As has been rightly said, AI is not one thing but a constantly evolving stack of technologies with ever-widening uses. At this stage, it would be counterproductive to subject this as yet nascent and constantly evolving technology to ex-ante regulation. India would be an outlier if AI is straitjacketed through ex-ante regulation.

Disclaimer: The views expressed are the author’s own and do not reflect the official policy or position of Financial Express.