Growth without work: The human cost of the AI revolution
Rest of World Jibu Elias ● Covered by 2 sources
AI is quietly gutting outsourcing jobs across India, the Philippines, Kenya, and Colombia—not just factory work, but transcription, IT, and design careers. The bigger shock: it's hitting educated, well-paid women hardest, and growth keeps rising while jobs don't.
Aakash thought he'd landed something stable. Twenty-five years old, a year into a Bengaluru transcription job, he'd asked his CEO point blank whether AI would come for his role. Five years away, he was told. It took less than one financial year. By March 2024, American clients were pulling contracts and routing audio through automated platforms instead of Indian typists. The 3 a.m. cafeteria rush disappeared. So did the recruitment team that once hired people like him.
What happened to Aakash is happening at scale, and it's not the industrial-robot story anyone braced for decades ago. This is cognitive work — the stuff that built India's outsourcing economy and gave millions a path into the middle class. Manila's transcriptionists, Nairobi's call-center staff, Colombia's customer-support teams: all watching the same script play out. Economists call these jobs "high exposure, low complementarity," meaning software doesn't make the human more useful, it just replaces them outright. Even people paid to train systems like Gemini are annotating tools that will eventually make their own annotation work obsolete.
The numbers are blunt. India's tech sector cut over 500,000 jobs between 2022 and April 2024, with 425,000 of those losses in 2023 alone. Net hiring across the country's top five IT firms in the first nine months of 2025 totaled 17 — not 17,000, seventeen. HCL's former CEO has floated a figure of 70% of IT roles at risk. Globally, trackers logged roughly 244,000 tech layoffs in 2025, with automation named repeatedly as a driver. Bioengineers show 84% exposure to AI disruption, mathematicians 80%, editors 72% — the fields once considered safest because they demanded judgment and creativity are now squarely in the blast radius.
There's a useful, if uncomfortable, historical parallel. In September 1945, 15,000 New York elevator operators walked out, and the city discovered how dependent it was on jobs nobody thought twice about. Manufacturers spent the next few years engineering those workers out of existence — automatic doors, alarm systems, self-service cabs — and by the 1970s the profession was gone. AI is doing something similar but faster and wider, hollowing out entire categories of work simultaneously rather than one trade at a time.
The distribution of pain is the real story here. Anthropic researchers matched actual AI usage data against U.S. labor projections and found the occupations most exposed to AI are disproportionately staffed by educated, well-paid women — precisely the workers told their degrees would insulate them. India's an outlier only because so many women are parked in agriculture, a sector AI hasn't touched yet. Meanwhile the IMF describes a Great Divergence: capital and productivity gains concentrate in economies rich enough to deploy the robots, while developing countries eat GDP declines and worsening terms of trade. The OECD estimates generative AI could add 6.4% to advanced-economy GDP while still displacing millions of workers within those same economies. Growth is happening. Paychecks and job security are not following it.
My take
Nobody in Silicon Valley is going to lose sleep over a Bengaluru transcriptionist or a Manila call-center rep, and that's exactly the problem — the people building these tools live nowhere near the fallout, so the fallout doesn't shape the roadmap. GDP growth stopped meaning job growth years ago, and pretending otherwise is just cover for treating layoffs as an efficiency win instead of a policy failure. If a society can add 6% to GDP while gutting the workforce that used to earn a living from that output, the metric everyone's celebrating is measuring the wrong thing.
Read more about this at: Rest of World
