AI and the Great White-Collar Recession

IN the 1990s, newspapers across India started putting out 'Page 3' stuff — colourful reports about the social life of the bold and the beautiful, the rich and the famous. Those were the early days of globalisation, when glamour and glitz had just started to capture the middle-class mind in India. Very soon, newspapers and magazines began to print supplements and pullouts that focused on the fabulous lives of socialites. In fact, to break into the world of the glamourous, it became important to break into Page 3.

Soon, socialites and social climbers began to cultivate glamour and entertainment reporters. This became an industry of its own — and a side hustle for Page 3 reporters who would charge money to cover a party or an event. It was an open secret, which, understandably, reached the ears of newspaper owners. One of them — a fine enterprising specimen — decided to join the party. So, s/he made the Page 3 pullout an entirely sponsored feature. You could throw a party, invite the usual suspects, hire a photographer, send the pictures to the newspaper, duly labelled with the right names, pay a fixed fee and get published in the paper.

Why am I telling you this? Because AI is a bit like Page 3. All of us have used it to speed up our work. It is very common to get a call from your boss at 9 pm, telling you to get a deck or a presentation ready by 9 am. Before AI platforms became really smart, you had to burn the midnight oil and get the document ready, only to see your boss take credit for it. Now, you can simply give a well-tailored prompt to ChatGPT, upload the output to ChatGPT for PowerPoint, and get a presentation out in 15 minutes.

Until a few months back, your boss ignored this subterfuge. S/he knew that you were using AI, but didn't want to confront you with it. Now, you are being asked to 'show' that you are using AI in your work. If you aren't, then you will be sent to an 'upskilling' workshop. If you still stubbornly refuse to make your work less 'human', you risk being served with a pink slip.

This is exactly what is happening across corporate India — indeed, across the world. Bosses have embraced the reality of AI, much like the newspaper owner had embraced the fact that people will pay to get onto Page 3. So, instead of allowing employees to reduce their work by using AI, they are using AI to reduce their employee strength.

To understand this better, think of a typical old-school corporate entity. It is structured like a pyramid. There are many employees at the bottom, with entry-level skills, who perform many mundane tasks. Above them are a smaller number of professionals — still young — who have more experience, have displayed greater skills and, therefore, do more intricate tasks. They get paid more, but don't have anyone reporting to them.

Right above this level come the middle-level managers — a slightly smaller body of leaders and supervisors who are in charge of small projects and have teams under them. This level is made up of young rising stars and older has-beens who didn't manage to rise to the next level. These people make schedules, supervise and track work output, and manage people under them, work on important but routine things such as compiling payroll data, or optimising employee leave applications. These are important roles that keep the corporate body alive, but they don't have strategic value.

Above these middle-level managers comes a layer of senior managers — people who supervise many middle-level managers, work on growing the business, plan cost reductions, long-term brand-building, and expansion strategies. At the top of this layer are the various Cs in the C-suite - COO, CFO, CMO, CTO, CHRO, and so on. And above all these layers, sits the CEO or the MD.

It has now become abundantly clear that AI — if used to its current limit — can handle not only the work of the first two levels of the pyramid but also many supervisory and administrative tasks of middle-level managers. Till a few months ago, enterprising employees in each of the bottom three levels, including mid-level ones, were using AI platforms to ease their work and speed it up. Now, the bosses are saying let AI do it directly — why do we need you?

That automatically means that a company needs far fewer entry-level employees, and even a significantly lower number of middle-level managers.

There is another reason for a chunk of middle-level managers becoming redundant. Imagine a company with a hundred entry-level and junior employees, supervised by 10 middle-level managers. Now, AI renders 40 of them redundant. The company is left with only 60 employees for supervision. Given that they, too, are required to use AI for their work, each middle-level manager will be expected to handle 15 subordinates. This means, instead of needing 10 middle-level managers, the company will need only four. Six middle-level managers become redundant.

This is happening across the world at an alarming pace. TeamLease Digital estimates that AI will put 20% to 40% of Indian middle-level managers at risk, across IT, outsourcing, finance, retail and manufacturing, in just one year. Many big companies — TCS is the most cited example — are sacking middle-level managers to make way for AI. In some cases, employers are sacking people, not because AI is doing a better job than humans, but often in anticipation of what is yet to come, even if AI hasn't delivered much for their own companies.

This is what some economists are calling the Great White-Collar Recession of 2025. The IMF has warned that AI could lead to long-term unemployment and economists agree that middle-class professionals will be the worst hit. If it continues like this, what started in 2025, is only going to get worse in 2026.

Aunindyo Chakravarty is a senior economic analyst.