New Delhi, Jan 29 (PTI) The Economic Survey has flagged "looming uncertainties" and "asymmetries" in the evolving AI landscape, drawing attention to issues from concentration of market power in hands of few large global frontier model builders, to labour dynamics, and all the way to questions around future of the Indian IT sector.

While India enters the AI era with notable strengths, access to cutting-edge compute infrastructure is limited and financial resources for large-scale model training remains scarce compared to global leaders, making pursuit of foundational models as the centrepiece of AI strategy "challenging", the Survey said.

Instead, it advocated a bottom-up approach to AI development saying it aligns more closely with economic realities. Value creation in AI need not be concentrated in a small number of frontier models or firms, it said speaking out in favour of application-specific, small models that are tailored to defined uses and sectoral needs.

Amid evolution of global AI structures, the survey highlighted a dangerous concentration of power where control over critical inputs - data, compute, and foundational models - is held by a few global entities.

On labour market implication, it said while the emerging evidence has provided some reassurance in the near term for economies such as India, there is no room for complacency, especially from a policymaker’s perspective.

"All in all, caution is still warranted as India attempts to solve the puzzle of AI and labour. This represents one of the most considerable looming uncertainties about the technology," the Survey said.

Uncertainity also extends into the evolving structures of the global AI ecosystem, it noted.

Control over critical inputs is increasingly concentrated, the Survey said noting this raises concerns about market power, technological dependence, and the resilience of supply chains.

"It also raises a substantial question about the future of India’s IT sector, as firms that once relied on India’s comparative advantage to handle a bulk of their work may no longer need to do so." This risks "hollowing out" India’s core value proposition if adaptation lags.

"If the country is to sustain its competitive edge in IT, a comprehensive evolution is necessary, one that takes full advantage of the potential embedded in AI development and deployment," it said.

Another source of uncertainty pertains to defining a regulatory approach to AI, it said, and added countries diverge in how they design their institutions to address AI-related challenges, reflecting differing priorities.

For advanced economies, their decisions will determine how AI can be leveraged to enhance productivity in the face of labour shortages and an ageing population.

"For India, the challenge is to govern AI in a manner that is sensitive to its economic realities. The choices that India’s own institutions make will play a central role in determining not only the pace of AI diffusion but also how its economic value is distributed across sectors and among people," it said.

The Survey highlighted the strategic necessity of India’s own AI solution and proposed practical steps that can be taken to evolve the ecosystem. This includes AI model and application development, human capital, and how governance needs to evolve, including a framework for incentivising data localisation.

It mentioned the global AI divide, defined by sharp distinction between frontier model development and application-led development.

The capability to design and train large foundational models remains highly concentrated in the hands of a few large firms, it said adding that these firms exercise significant control over the market and exert high demand pressures on the resources necessary for AI, allowing them to erect high barriers to entry.

Add to this, the export restrictions imposed on the most advanced processors required for scaling up frontier model development, and the task ahead of India becomes extremely challenging.

"This creates a fundamental asymmetry: most countries may end up participating in AI primarily as users, while a few will shape the technology's trajectory, standards, cultural leanings, and pricing. Attempting to close this gap would involve prohibitive fiscal costs towards what is increasingly becoming an unsustainable approach to AI development," it said.

The trade-off, therefore, is between using scarce resources to chase frontier-scale models or deploying those resources more effectively towards domain-specific AI systems aligned with domestic economic priorities.

For India, the policy challenge, is not whether to adopt AI, but how to pace its diffusion so that labour augmentation can be facilitated.

"Rapid, uncalibrated deployment of AI may boost output but risks displacing segments of the workforce faster than the economy can reabsorb them. Conversely, delaying adoption to protect jobs may risk locking firms into a low productivity equilibrium," it said.

Another asymmetry lies in the ownership and governance of AI models.

India’s challenge is to strike a balance between openness and stewardship, leveraging shared innovation while ensuring that the economic value created from domestic data and intellectual property accrues within India rather than being captured abroad. PTI MBI MBI ANU ANU