1 hour ago
5
ARTICLE AD BOX
Every CEO in India today is staring at the same irresistible choice: deploy AI, reduce headcount, boost productivity and protect margins. PwC's 2026 AI Jobs Barometer confirms it: companies most exposed to AI are clocking 40% higher productivity growth than the least exposed.
At an individual company level, the math makes perfect sense. The problem starts the moment every CEO does the same math, at the same time, across the same economy. This is a classic prisoner's dilemma - a situation where the most rational individual choice produces the worst collective outcome. Each company, acting purely in self-interest, makes the obviously correct choice to automate. But when every company makes that choice simultaneously, the workers being replaced don't just disappear from payrolls but also from the consumer base.
Fewer paychecks mean less spending, and less spending morphs into weaker demand for everyone, including the companies that automated first to protect their margins.No one is untouched by it. Every B2C company is ultimately selling to a person whose income comes from somebody's payroll. And every B2B company ultimately sells to a B2C company that depends on the same person. Because there’s no individual way to escape a problem that is being caused by everyone together, it becomes a trap.
AI does pay off for the company deploying it. But the payoff also erodes the demand base every company, automated or not, eventually depends on. This brings us to a very critical yet seldom discussed outcome of large-scale AI adoption – its impact on aggregate demand. Going beyond the current conversations around AI ethics, bias, and job displacement, we also need to invest time and resources to analyse how widespread automation could affect consumer demand in the long run.
In fact, the impact is already playing out in the market and is reflected in numbers.Hello AI, Bye-Bye consumer demand2026 has been brutal for the tech sector: over 1.16 lakh jobs lost globally in the first five months alone, the steepest cut in nearly two years. In India, Oracle, Amazon, TCS, Meta, Flipkart and Livspace have all trimmed workforces, with automation cited as a driver. Researchers at Marcellus find that white-collar job growth in India has slowed to just 1% annually between 2023 and 2025, and quote AI as a reason.
NASSCOM-BCG-NITI Aayog report projects that up to 2 million roles, nearly a quarter of the tech and customer-experience workforce, could disappear by 2031 under a high-disruption AI scenario. Ironically, this cohort is also the engine of India's rising middle class, which drives roughly 60% of GDP through consumption.
When a meaningful slice of that middle class spends months and years unemployed, it slows down spending and weakens demand.
For example, home sales across India's top cities fell 13% in the first quarter of 2026, according to data from Knight Frank India, a real estate consultancy, and economists cite IT-sector layoffs in Bengaluru, Hyderabad and Pune as a reason. Earlier it was Man vs Machine, now AI vs ManEvery prior wave of technology has destroyed jobs and eventually created more. The power loom wiped out 3-4 million handweavers and artisans in the late 18th century, but it also built the factory system and a textile industry that today employs close to 90 million people globally.
Office automation in the 1980s erased millions of clerical roles, but it midwifed an IT industry that, at its peak, employed 18-20 million people worldwide.
A widely cited 2011 McKinsey estimate even put it in a clean ratio: the internet created 2.4 new jobs for everyone it destroyed.AI has a different script. It isn't replacing an older, slower technology with a faster one; it's replacing human input and presence across knowledge work and routine work simultaneously. Sam Altman has put a number on this himself, suggesting that somewhere between 30% and 40% of today's economic tasks could be handled by AI within the next few years. With no other large industry standing by to absorb the displaced workforce, one wonders with trepidation how this will play out economically in the long term. That’s why this calls for an urgent and deliberate discussion from all stakeholders involved: government, regulators, industry bodies, IITs, IIMs, and think tanks.
Everybody needs to pitch in and think of ways to build effective pathways for labour absorption, reskilling and creation of new economic opportunities.
Building guardrails: what AI can learn from climate action building We have, in fact, solved a structurally similar problem before – climate change. About 25-30 years ago, every nation in the world was thinking about its self-interest and gro wth when it came to reducing carbon emissions. No individual country or organisation had any incentive to reduce emissions on their own while competitors chased growth unfettered.
It was a prisoner’s dilemma at planetary scale. The response wasn't left to individual goodwill. The Kyoto Protocol (1997) and later the Paris Agreement (2015) forced governments to commit to binding targets. Carbon pricing, emissions disclosure norms, and ESG reporting forced corporates to internalise a cost that the market wasn't pricing in on its own. AI's labour and demand shock need a similar treatment, urgentlyLeft to pure market incentives, every company will keep making the individually correct call, and the economy will keep paying the collective bill, until the demand-side damage spirals into a crisis.
Investors have a role here too. They need to back AI deployment that compounds productivity without hollowing out the market, their own portfolio companies will need five years from now.Technological disruptions are inevitable, but the chaos they create can be avoided or mitigated. By developing a framework to govern the AI-led transition, we can ensure a stronger, more resilient economy and a safer future for the 1.2 crore young Indians who enter our workforce every year. Devendra Agrawal, founder, Dexter Capital Advisors
English (US) ·