The Centre for Responsible AI at IIT Madras (CeRAI-IITM) has released a White Paper arguing that artificial intelligence is reshaping India's labour market through task-level transformation rather than wholesale job elimination — and calling on employers, workers, and policymakers to stop treating every instance of AI-driven change as a single, uniform story.

Key Argument: A Taxonomy, Not a Binary

The paper's central thesis challenges the dominant public narrative around AI and employment. 'Public debate on AI and Indian employment tends to compress a complicated phenomenon into a single binary: Will AI take a given job, or not?' the document states. 'This paper argues that the accurate picture is a taxonomy rather than a binary. AI rarely replaces a job wholesale; more often it automates a bundle of tasks within a job, and the size of that bundle varies with task structure, sector, regulatory setting, and, distinctively in India's case, the economics of AI access itself.'

To address this complexity, the researchers propose a four-level analytical framework spanning task, job, occupation, and labour-market outcome. Indian jobs are classified into two broad regimes: those where AI has automated most of the skill bundle — leaving the human largely as a validator — and those where the human remains the primary driver and AI functions as a productivity tool. Each regime is further broken down into policy-relevant sub-types.

Sectors Examined and What They Reveal

Drawing on evidence from business process outsourcing (BPO), manufacturing, MSMEs, sanitation, financial services, law, and software engineering, the researchers found that every job studied had to be reimagined — through task-level augmentation, task reallocation, or the creation of genuinely new roles built around operating AI systems — rather than simply eliminated or left unchanged.

The paper identifies five recurring reasons why organisations are adopting AI: cutting costs, reaching capabilities available only through frontier-grade AI, meeting regulatory requirements, improving safety by removing humans from hazardous work, and correcting service quality after earlier failed automation attempts. Notably, each reason, according to the researchers, demands a distinct policy response rather than a one-size-fits-all approach.

Policy Recommendations: Three-Pronged Approach

The paper closes with a distributional analysis of who bears the costs of workplace reimagining and outlines three policy priorities: continuing to encourage AI diffusion across the economy, investing in India-specific labour-market research, and pairing AI adoption with structured worker-transition pathways. 'Each reason calls for a different response. Cost-driven automation needs support for workers in transition, and vigilance against over-automation. Capability-driven automation needs investment to make frontier-grade AI more affordable, not just more available,' the document states.

The researchers further warn that regulation-driven automation needs clear liability rules set up in advance, safety-driven automation should not be contingent on cost savings, and quality-driven automation rollbacks need 'humility about how far automation can be pushed before it breaks.'

A Long-Term Warning on Expertise and the Pipeline Problem

One of the paper's more pointed observations concerns the reproduction of expertise. If AI absorbs much of the entry-level work through which people have historically acquired domain knowledge and judgment, the researchers ask, how will organisations develop the next generation of human reviewers, supervisors, and experts? This so-called 'pipeline problem' is flagged as a long-term structural risk that current policy frameworks have yet to address.

The paper also notes that as AI evolves from assisting on individual tasks to executing bounded decisions and entire workflows, the unit of change may itself shift from the task to the workflow level — making continuous reassessment essential.

What Comes Next

The White Paper is described as the first step in a broader research programme. Subsequent work under CeRAI-IITM will include systematic evidence reviews, sector- and work-profile analysis, and primary field research with employers and workers across areas such as education and healthcare. Future studies will also connect with CeRAI's ongoing research on AI governance, responsible AI principles, and a public perception study conducted with CMC. 'Multiple AI stories are unfolding across the Indian labour market at once, and the right response depends on knowing which one is in front of you,' the researchers concluded.