AI to reshape India's workforce? 197 million workers face automation risk - Mint

TEXT START: Nearly 197 million Indian workers are employed in occupations with a high share of routine tasks that could be automated, while about 46 million work in jobs where artificial intelligence (AI) could enhance productivity, according to a new study based on India’s labour market data.

The Dissection

The article converts a structural rupture into a policy brief. It identifies a large population exposed to automation, then softens the implication by dividing workers into “at risk” and “benefiting” groups and prescribing reskilling, AI literacy, and expansion of supposedly human-essential sectors.

The useful contribution is the attempt to measure India’s exposure through domestic labour data rather than imported estimates. But the article never clearly reconciles its figures: 197 million workers are described as occupying highly routine jobs, while only 11.4% are said to face substitution risk; 46 million supposedly benefit, equating to 7.6%. Exposure, task automation, productivity assistance, and actual displacement are treated as adjacent concepts without a clear denominator or causal chain.

The Core Fallacy

It assumes that “complementary” occupations remain durable human domains. Healthcare, education, management, science, engineering, and care work may initially use AI as an assistant. Under Cognitive Automation Dominance, assistance becomes substitution once systems outperform humans in diagnosis, instruction, planning, documentation, coordination, and judgement at lower cost.

The second fallacy is treating reskilling as a solution to a structural ownership problem. Training can move some workers into narrower supervisory, verification, maintenance, or liability-bearing roles. It cannot manufacture enough economically necessary human labour when AI reduces the quantity of labour required. The article offers a ladder into a shrinking structure and calls it a bridge.

Hidden Assumptions

• Productivity gains will translate into broad employment rather than accrue primarily to owners of AI, platforms, data, and infrastructure.
• Human interaction will remain economically necessary instead of merely socially preferred.
• States can coordinate a stable human-only economic perimeter despite competitive pressure to automate.
• Reskilling capacity will scale faster than displacement.
• “Higher-productivity employment” will exist in sufficient volume for workers leaving routine occupations.
• Regional differences represent different development paths rather than different speeds of exposure.
• The social system can preserve mass consumption after productive participation has collapsed without triggering a deeper legitimacy crisis.

None of these assumptions is established by the supplied text.

Social Function

Primarily transition management and ideological anesthetic, with a substantial partial truth. The report accurately detects uneven exposure and correctly warns that preparation must precede visible job losses. But its remedies redirect attention from ownership and power toward worker adaptation. “Reskilling” becomes the respectable language for rationing access to the remaining useful roles while leaving the AI capital structure intact.

The state-level comparison is also misleading if read as a map of safety. Kerala, Haryana, Maharashtra, and Tamil Nadu may possess more workers in currently complementary occupations. That is a lag advantage, not an escape route. Uttar Pradesh, Bihar, Jharkhand, and Madhya Pradesh face earlier weakness, but the better-positioned states remain exposed when AI advances from routine execution into expertise, coordination, and judgement.

The Verdict

This is a competent exposure report wrapped in a survivability fantasy. Its strongest fact is that India has a massive automation risk concentrated among workers with limited bargaining power. Its fatal omission is that the 46 million potential beneficiaries cannot replace the productive function of the 197 million exposed workers, especially if the gains are privately captured.

The article describes the opening phase of productive participation collapse, then proposes training as if the problem were a skills mismatch. It is not. The problem is that AI progressively severs the wage-to-consumption circuit. India can delay the break through lagging infrastructure, regulation, public employment, and labour-intensive sectors. It cannot reverse the underlying mechanics without controlling the ownership of AI capital and the distribution of its output.