New Delhi: A new study by researchers at Stanford University and the United States’ National Bureau of Economic Research offers some of the clearest evidence yet that generative AI is beginning to reshape labour markets. The troubling finding is that young workers are experiencing the impact first — a signal that holds particular significance for India, given its young and rapidly expanding workforce.
The US study is based on high-frequency payroll data covering more than 25 million workers. By tracking monthly employment patterns through September 2025, the researchers were able to observe how hiring shifted after generative AI tools such as ChatGPT became widely adopted in late 2022.
Their key finding is stark. In occupations most exposed to AI — including software development and customer support — employment among workers aged 22 to 25 has fallen sharply. After controlling for firm-level shocks, early-career workers in the most AI-exposed roles experienced roughly a 16% relative decline in employment. Older, more experienced workers in the same occupations did not see similar declines; in many cases, their employment remained stable or even increased.
The adjustment is visible primarily in hiring, not wages. Compensation has not fallen dramatically across exposure categories. Instead, firms appear to be slowing or reducing junior recruitment rather than cutting pay. In simple terms, the entry gate is narrowing.
This pattern reflects the kinds of tasks generative AI performs well. Current systems excel at structured, administrative and rule-based activities. They can draft routine emails, summarize documents, generate standard code, answer common customer queries, extract information from forms and prepare first drafts of reports. These tasks rely on codified knowledge — information that is written down, documented and governed by explicit rules.
Traditionally, such responsibilities have been handled by entry-level employees. Fresh graduates in IT firms write test scripts and documentation. Junior analysts prepare spreadsheets and presentations. Customer support associates handle standardized queries before escalation. These roles form the first rung of the corporate ladder.
AI can now perform many of these functions quickly and at low marginal cost, weakening the economic case for hiring large cohorts of junior staff to carry out routine work.
At the same time, AI continues to struggle with nuance, contextual judgment and tacit knowledge. It does not genuinely understand organizational politics, client relationships, informal norms or the unwritten rules that shape decision-making. It cannot easily replicate the judgment built through years of navigating ambiguous and complex situations.
This distinction between codified and tacit knowledge helps explain the age divide in employment outcomes. Young workers tend to contribute more codified knowledge — what they learned in classrooms, technical frameworks and formal methods. Experienced professionals bring tacit knowledge: pattern recognition, situational awareness, institutional memory and accumulated judgment.
If AI substitutes for codified tasks but complements tacit skills, early-career workers are naturally more exposed. Experience becomes a protective buffer.
The concern, however, is long term. Tacit knowledge is not automatic; it is accumulated gradually, often through performing routine tasks under supervision. Entry-level roles function as apprenticeships, allowing young professionals to convert theoretical knowledge into lived expertise.
If companies reduce hiring at the base of the pyramid because AI handles routine work, fewer young workers will get the opportunity to build that experience. The protective buffer may never fully form for the next generation.
The India Impact
For India, this dynamic intersects with existing structural pressures. The Periodic Labour Force Survey released by the Ministry of Statistics and Programme Implementation shows that while overall employment indicators have improved in recent years, unemployment among educated youth remains significantly higher than among older cohorts. The mismatch between graduate output and the availability of quality entry-level jobs has already been a policy concern. If AI reduces demand for routine white-collar roles, that pressure could intensify.
Generative AI directly touches many tasks at the base of India’s technology and services pyramid: code generation, documentation, ticket resolution, compliance processing and standardized reporting. Indian firms are already deploying AI copilots across development and support functions. Recent annual reports from leading IT services companies show moderation in campus hiring and greater emphasis on reskilling and lateral recruitment. While global demand cycles and macroeconomic conditions play a role, AI-driven productivity gains are increasingly part of the equation.
The risk is not immediate mass unemployment. Rather, it is slower entry into structured, skill-building careers. If securing the first job becomes more difficult, the entire career trajectory can shift. Fewer opportunities at the start may translate into thinner leadership pipelines later.
If companies scale back hiring at the base of the pyramid as AI absorbs routine tasks, fewer young professionals will gain the hands-on experience that shapes real expertise. Without those early opportunities, the protective cushion of judgment and tacit knowledge may never fully develop for the next generation.
Researchers who study career development often compare professional growth to a tree. Early in a career, the trunk forms through shared foundational tasks. Over time, branches grow as workers specialize and deepen expertise. If the trunk does not develop fully — if early experience is truncated — the later branches may be weaker or fewer.
For Indian firms, this carries strategic implications. The country’s competitive advantage has long rested on its ability to train large numbers of young professionals and progressively move them up the value chain. If AI reduces the need for junior staff while increasing demand for higher-order judgment and client-facing capabilities, organizations will need to rethink how they cultivate talent.
One response may be to redesign entry-level roles rather than eliminate them. Instead of focusing on routine output, junior positions could emphasize supervised problem-solving, exposure to real client contexts and structured learning alongside AI tools. Used thoughtfully, AI could accelerate skill acquisition rather than replace the learning process.
Educational institutions may also need to adapt. Producing more graduates with standardized technical knowledge may not be sufficient. Greater emphasis on critical thinking, communication, domain expertise and adaptability could help young workers move more quickly toward roles that rely on skills AI struggles to replicate.
Technology transitions have historically created new forms of employment even as they disrupted existing ones. But they rarely do so evenly. The US payroll evidence shows that generative AI’s impact is already measurable, and that early-career hiring in highly exposed occupations has weakened.
India adds millions of graduates to its labour pool each year, many of whom depend on entry-level roles in IT services, business process management and corporate back offices as their first step into the formal economy. According to NASSCOM, India’s technology industry employed more than five million professionals in FY2024, with a staffing model built on a broad base of freshers supporting a narrower layer of experienced managers and architects. If the base of that pyramid weakens under AI pressure, the long-term effects could be significant.
For India, the lesson is clear. Experience is becoming more valuable because it embodies capabilities that AI cannot easily codify. Yet experience develops only when young people are given the chance to work, learn and make mistakes. If automation narrows the entry gate too sharply, India risks weakening the very pipeline of expertise that has powered its rise in the global services economy.