AI job-loss fears for IT firms 'hogwash', disaster marketing by large capital: Nandan Nilekani

TEXT START: Infosys co-founder Nandan Nilekani on Tuesday dismissed fears that artificial intelligence would wipe out jobs in India’s IT services sector as “hogwash” and “disaster marketing to suit the interests of large capital”, saying demand for such firms is higher than ever.

The Dissection

This is an incumbent IT-services defense brief disguised as an economic forecast. Nilekani shifts the question from whether AI destroys labor demand to whether enterprises will need help implementing AI. That substitution is the central evasion.

More transformation projects can produce more revenue for Infosys while requiring fewer employees per project. A busier intermediary is not proof of a healthier labor market. It may simply be the contractor installing the machinery that eliminates its own workforce.

The article also converts delay into destiny. Legacy systems, power structures, and slow adoption can postpone displacement. They do not repeal it.

The Core Fallacy

The text confuses increased demand for AI-related services with increased demand for human workers.

AI can expand the market for transformation while compressing the labor required to perform transformation. Coding is already identified as automatable; requirements analysis, coordination, governance, testing, monitoring, documentation, and change management are not metaphysically human. They are cognitive workflows. Once AI becomes capable enough, they become targets too.

The claim that technology is only 30 percent while governance and transformation are 70 percent assumes that the 70 percent remains labor-intensive. Under the Discontinuity Thesis, that is precisely what fails. Complexity is not a permanent moat. It is an inventory of processes awaiting decomposition and automation.

“AI-proof jobs” is equally dishonest language. Jobs protected by temporary fragmentation, low adoption, or social friction are not proof against AI. They are lag pockets.

Hidden Assumptions

• AI adoption will remain slower than competitive pressure can tolerate.
• New AI work will create more jobs than it displaces.
• Governance, collaboration, and change management require permanently large human workforces.
• AI agents will remain too unreliable for automated verification and control systems to replace much of this labor.
• Small businesses will use AI to hire additional workers rather than produce more with fewer.
• Productivity gains will translate into broad wage employment instead of higher margins and concentrated ownership.
• Displaced IT workers can reskill into new roles faster than those roles are automated.
• Demand growth for Indian IT services will outpace the productivity growth that reduces staffing needs.
• Millions of small firms can collectively recreate the mass employment and consumption circuit of postwar capitalism.

None of these assumptions is demonstrated. Most are simply smuggled in through the word “transformation.”

Social Function

This is elite self-exoneration and transition management wrapped around a partial truth.

The partial truth is that enterprise adoption is slow and that transitional demand for consultants, integrators, auditors, and implementers will rise. The ideological anesthetic is the claim that transitional demand guarantees durable mass employment.

The narrative reassures workers, policymakers, and investors that the incumbent system is adapting rather than dying. It also protects the commercial position of firms like Infosys: if AI threatens their labor-arbitrage model, redefine the threat as a larger consulting opportunity.

The small-business argument performs the same function at national scale. It replaces evidence of durable employment with a visualization of millions of hypothetical firms hiring “a couple of people each.” That is not an economic mechanism. It is a hopeful sketch standing in for demand, capital allocation, survival rates, and net headcount.

The Verdict

Nilekani is right about the lag and wrong about the destination. Infosys may receive more transformation work while requiring fewer workers to deliver it. AI deployment can expand before productive participation contracts; that is the transition phase, not a refutation of the Discontinuity Thesis.

The article mistakes the scaffolding around automation for an escape from automation. Governance becomes software. Coordination becomes agents. Verification becomes automated control. Small firms become leaner, not necessarily more labor-intensive.

This is not an argument that the postwar employment circuit survives. It is a polished lullaby from a company positioned to profit while that circuit is dismantled.