India Built the World's Back Office. AI Just Fired It.

The Tables Have Turned. Are You Ready?

For years, Indian coders were the world’s back office. Now AI is doing the same to them that they did to Western workers — and the biggest consulting firms in the world are quietly cashing in.

 

BY THE NUMBERS

20,000+ TCS jobs cut in 2025  ·  586 tech workers laid off globally every single day  ·  64% of Indian IT firms now using generative AI tools  ·  ₹2.4 crore charged by Deloitte for a report that was mostly fake

 

PART 1

The Mirror Moment India Didn’t Expect

Remember the anger in America and Europe when Indian software companies “stole” their jobs? H-1B visas, outsourcing, off-shoring — these were dirty words in the West for two decades. The argument was simple: why pay a US developer $120 an hour when you could get an equally smart engineer in Bengaluru for a fraction of that?

Now the tables have turned — and the new low-cost replacement isn’t a person in another country. It’s a software tool that costs $20 a month.

AI coding assistants can now write boilerplate code, debug errors, generate test cases, write documentation, and build basic features — in seconds. The tasks that used to justify paying a developer ₹2–3 lakh per month are increasingly being done by AI tools that companies pay a flat monthly subscription for. No visa. No management. No attrition.

“Entry-level salaries at India’s top five IT firms increased by less than 10% over 15 years, while the cost of education, food, and housing rose exponentially.”
 — Maheshwer Peri, founder of Careers360, career research study, 2025

India’s IT industry was built on a single competitive advantage: cheap, skilled, English-speaking labour at scale. That advantage is now being undercut — not by cheaper humans elsewhere — but by machines that don’t sleep, don’t ask for raises, and don’t resign to join a startup.

TCS, India’s largest private employer with over 5 lakh employees, announced more than 20,000 job cuts in 2025 as part of what it called an “AI-first delivery model.” Infosys and Wipro together shed another 25,000–30,000 positions. The reasons given were sanitised corporate language: “skill mismatch,” “limited deployment opportunities.” But everyone in the industry knew what it really meant.

 

PART 2

The Deloitte Scandal That Should Have Been Front-Page News

Now here is where it gets darker — and more absurd. Because while employees are being replaced by AI to save money, the companies doing the replacing are using AI carelessly, passing off sloppy AI output as expert work, and charging governments crores for it.

Let’s talk about Deloitte.

CONFIRMED SCANDAL · AUSTRALIA · 2025

Deloitte Charged ₹2.4 Crore for a Report Full of Fake Citations

The Australian government hired Deloitte — one of the world’s “Big Four” consulting giants — to review its welfare compliance system. Deloitte charged AU$4.39 lakh (about ₹2.4 crore) for a 237-page report. When a university researcher named Dr Chris Rudge sat down to read it, he spotted something immediately wrong: the report cited books, papers, and court judgments that simply did not exist. One citation named a professor as the author of a book she had never written. Another quoted a federal court judge — except the quote was entirely made up. Deloitte had used an AI tool (Microsoft’s Azure OpenAI GPT-4o) to write the report — and never told the client. It was forced to issue a corrected version and refund part of its fee. Australian Senator Barbara Pocock called it out: “the kinds of things a first-year university student would be in deep trouble for.”

“I instantaneously knew it was either hallucinated by AI or the world’s best kept secret — I’d never heard of the book and it sounded preposterous.”
 — Dr Chris Rudge, Sydney University researcher, on discovering the fake citation

But wait — that was just Australia. Within weeks, the same story appeared in Canada.

SECOND OFFENCE · CANADA · 2025

Deloitte Did It Again — This Time for ₹10+ Crore

A 526-page Deloitte report for the Canadian province of Newfoundland and Labrador — on healthcare — was found to contain similar AI-generated errors. The Canadian government had paid nearly $1.6 million USD (about ₹13 crore) for it. The report cited a nursing professor as the co-author of an academic paper that did not exist. She confirmed she had never written it. As of the time of reporting, the report was still live on the government’s website. No refund had been announced.

Think about what this means. One of the most prestigious consulting firms in the world — a company that charges governments millions — was quietly using AI to generate content, not telling anyone, and submitting hallucinated citations as expert research. They were not “assisting” with the report. They were outsourcing it entirely to a chatbot and charging human expert rates for it.

 

PART 3

Deloitte Isn’t Alone — This Is Happening Everywhere

The Deloitte cases are symptoms of a wider epidemic. Across industries, people are passing off AI-generated work as their own — sometimes by accident, sometimes deliberately.

LEGAL · USA · 2023

Two Lawyers Fined for Submitting Fake AI Case Citations to Court

Two US lawyers used ChatGPT to draft a legal brief in a personal injury case. The AI invented six court cases that had never happened. The lawyers submitted them to a federal court without checking. When the judge asked for the actual documents, they couldn’t produce them — because the cases didn’t exist. Both lawyers were fined and publicly reprimanded. This was one of the first high-profile cases that showed the world: AI doesn’t just make mistakes. It makes confident, well-formatted, completely wrong mistakes.

CONSULTING · GLOBAL · ONGOING

Big Four Firms Racing to Replace Junior Staff With AI

In June 2024, the UK’s Financial Reporting Council formally warned that the Big Four firms — Deloitte, PwC, EY, KPMG — were failing to monitor how AI tools were affecting the quality of their audits. These are the same firms that have invested hundreds of millions into AI to cut costs — specifically by reducing junior staff. The irony is bitter: they fire the humans who would have fact-checked the AI’s work, then submit the AI’s unchecked work to clients at premium prices.

 

PART 4

What This Means for the Average Indian Professional

Let’s be honest. The disruption is real. The numbers are stark. In 2025 alone, over 2.45 lakh tech workers worldwide were laid off — about 674 every single day. In 2026, the pace has only accelerated. The tasks most at risk are the ones that form the backbone of India’s IT industry: routine software testing, data entry, basic code generation, report writing, documentation, customer support scripting.

These are not “low-skill” jobs. These are jobs that took years of education and preparation. But they are repeatable, structured, and therefore — automatable.

“A wave of suicides among Indian tech workers — 227 reported cases between 2017 and 2025 — points to how deeply the pressure of this industry runs. “It reflects a deeply rooted problem in the corporate culture of the IT sector.””
 — Rest of World investigation, February 2026

But here is what the doomsayers miss: the Deloitte scandal proves that AI alone is not enough. AI hallucinates. It invents citations. It can’t go to court and defend its reasoning. It can’t take responsibility. It can’t understand the nuance of a local law, a cultural context, or a client relationship.

What AI cannot replace — at least not yet — is the person who knows enough to catch what the AI got wrong.

 

PART 5

So Who Wins? Who Gets Left Behind?

Here is the new reality, stated plainly:

→    If your job is to produce the same output, following the same steps, every day — AI will replace it. Not tomorrow, but sooner than you think.

→    If your job requires judgment, context, verification, client trust, or creative problem-solving — AI is your assistant, not your replacement.

→    Companies that use AI carelessly (like Deloitte did) will embarrass themselves publicly and lose client trust. The ones that pair AI with strong human oversight will win.

→    The person who knows how to use AI tools, and knows their limits, is worth far more than either a person who ignores AI or an AI with no human checking its work.

→    The bubble that has burst is not talent itself — it is the illusion that volume of output equals value. Writing 10,000 lines of boilerplate code is no longer impressive. Writing 100 lines that solve a genuinely hard problem still is.

 

The Indian IT industry became powerful by being better at cost-efficient execution than anyone else. That edge is gone. The new edge is understanding what AI can and cannot do — and being the human in the room who makes the difference.

Companies are not going to stop hiring people. But they will hire fewer people who do more. And those people will need skills that go beyond writing code: the ability to architect systems, question assumptions, communicate with clients, verify AI outputs, and take ownership of decisions.

The Deloitte scandal is a warning shot. Not just at Deloitte. At every firm that thinks it can swap human expertise for AI output, slap a logo on it, and charge premium rates. The clients are waking up. The researchers are reading the footnotes. The senators are asking for full refunds.

The era of being paid for showing up and doing routine work is ending. The era of being paid for judgment, creativity, and the courage to say “the AI got this wrong” — that era is just beginning.

 

— End —