AI could trigger big shift in labour markets
The survey highlighted several implementation challenges, including significant infrastructure requirements beyond traditional needs.
India’s labour markets are already shifting in anticipation of artificial intelligence (AI) despite the technology’s current limitations, the Economic Survey warned on Friday, calling for institutional frameworks to manage the transition and provide safety nets for affected workers.

The survey emphasised that India faces a magnified challenge due to its population size and relatively low per capita income, particularly as a services-led economy. It cautioned policymakers against dismissing concerns about AI’s adverse impact as “exaggerated” or “overestimated,” especially regarding its effect on entry-level jobs and potential economic disparities.
“If companies do not optimise the introduction of AI over a longer horizon and handle it with sensitivity, the demand for policy intervention and fiscal resources to compensate will be irresistible,” the survey noted, calling for greater corporate social responsibility.
The survey raised fundamental questions about AI’s necessity, noting that while many AI and machine learning tools have been in use for years, public recognition surged only after OpenAI’s ChatGPT release. It particularly highlighted AI’s potential impact on knowledge-based and creative jobs, though noting its practical limitations due to the high degree of human input required.
To address these challenges, the survey recommended a tripartite “compact” between the government, private sector, and academia to ensure widespread distribution of AI benefits. It outlined the need for three types of institutions: enabling institutions for worker skilling, insuring institutions for financial safety nets, and stewarding institutions for policy development.
The survey highlighted several implementation challenges, including significant infrastructure requirements beyond traditional needs. These include specialised components like chips and data centres, along with substantial data requirements. It warned that AI models need carefully cleaned data to avoid biases, though the cleaning process itself could introduce new biases.
Power consumption emerged as a critical concern. The survey cited Bloomberg data showing that power consumption by AI data centres is expected to reach 1,580 terawatt-hours, equivalent to India’s total electricity consumption. This increasing power demand could lead to greater reliance on natural gas-based electricity, potentially increasing emissions. The survey noted that this new source of emissions does not seem to have been factored in while making policies.
The survey also noted that water usage for cooling data centres, which currently uses drinking water, needs to be addressed along with the efficient use of land and rare earth minerals.
In the education sector, the survey suggested several AI applications, including teacher development and personalised student tutoring. It recommended using AI to automate tasks like lesson planning and assessment development, allowing teachers to focus on instruction and mentoring. The technology could also facilitate personalised learning experiences and career counselling.
For the banking sector, where AI is already being used for credit underwriting, regulatory capital planning, and fraud detection, the survey emphasised the need for robust governance frameworks. It identified several risks, including: the technology’s “black box” nature making reliability assessment difficult, challenges in establishing accountability and liability, inadequate human oversight and potential loss of expertise, cybersecurity vulnerabilities, potential for malicious uses such as synthetic identity frauds, market manipulation risks and increased dependence on third-party service providers.
While noting that widespread AI adoption might be gradual due to current limitations, the survey stressed the importance of establishing institutions now to prepare for the transition. It particularly emphasised the need to help workers in low-skill and low-value-added services, who remain particularly vulnerable to AI displacement.
The survey concluded that existing regulations must be amended to ensure AI’s use balances innovation with accountability and transparency, while remaining aligned with societal values. It emphasised that current AI limitations, including its tendency to “hallucinate” and its unreliability, must be recognised while developing these frameworks.
Union IT minister Ashwini Vaishnaw’s recent comments on the government’s plans to fund indigenous foundation AI models were noted, along with his suggestion that power consumption could decrease with algorithmic efficiencies, citing DeepSeek’s evolution as evidence of potentially lower-cost AI development.
