On Friday morning as I scanned headlines, one jumped out at me. The Guardian had a report out a couple of days ago on a new trend, the one where most employees were finding that the use of AI was not leading to productivity gains. Instead, it was leading to more AI slop at work, forcing employees to spend time fixing it. In the same report, there was mention of a recent survey. It was this survey that caught my attention. It found that most employees believed AI saved them no time. But in contrast, around 92 per cent c-suite executives believed it made them more productive.
It is a rather perplexing finding. After all, if AI is a productive and magical tool, it should work for everyone equally well. So, why such a huge contrast? Why does a certain set of people find it magical, whereas others believe it does nothing but produce slop? Turns out AI is indeed magical but not if you are poor.
As I went through the Guardian report, my thoughts turned to a work colleague. He leads a team of journalists and he is the de facto AI evangelist on the floor I work on. I mean he is deep in the world of AI and his experience is as magical as one can hope with various AI tools. But his enthusiasm about the latest Claude is not shared, or at least not to the same degree, by most of us.
Now, there could be several reasons for this. But one I can infer, rather lazily I must say, is that unlike most of us he is paying for the best AI tools. While we are still stuck with the free ChatGPT or DeepSeek, he is paying Rs 2,000 every month for Claude Opus. And he has been doing that for nearly two years now. He was earlier on the pro plan of ChatGPT but once it became clear in recent weeks that Claude is the best one can get, he moved to that.
Rs 2,000 per month — or Rs 24,000 a year — is not a lot of money for many white-collar workers. But it is also not exactly cheap. In fact, it is one of the steepest subscriptions you can buy in India. ChatGPT and Google Gemini might be a step behind Claude, but they too largely cost the same. Rs 2,000 seems to be the entry point to usable — and in some sense magical — AI experience.
In this lies the rub. AI models like Claude have become magical in many ways. But unless you pay for them, they are not going to be very useful. Not only is the free access to these tools limited to older or smaller AI models, even the usage limits are minuscule. If you want to really push the AI in creating or doing something that would make the cut, something that would be closer to what you as a human can create, then you have to use it a lot and be persistent with it. That is not possible without a properly paid account.
For example, our de facto AI champion has been at it for the last two years, training his AI model to think in a certain way, trying to mould it to his workflow and his taste. None of that is possible if you cannot shell out at least Rs 2,000 per month, although if you can go a step further and get a Pro Plus or Max plans, that would be even better. I know one more person in my office who too finds AI magical. Or at least one of the AI tools — Seedance, an AI model for generating videos. Guess how much this person spends on Seedance? Rs 16,000 per month.
So far this has been the story of AI. The best AI tools, the kind that are needed to recreate magic, currently are way too expensive for most people in India. Even with paid AI plans, there are usage limits that make it difficult to get the best output for almost everyone except the well-off white-collar workers. These are the people who lead teams and wear suits.
Is AI better at work rich people do?
Access to the best AI tools is one aspect of the story, just one part of the rich and the poor divide that AI is highlighting. Another aspect is what are you using it for? Rich people tend to use AI differently, compared to poor people. Just the way employees on an office floor will use AI differently compared to how Claude would be used in the cabin of a c-suite executive.
Here is an example: an employee might use the AI to write an email or prepare a sales report. A tool like Claude or Gemini would do a good enough job, but it is unlikely to make the employee say wow. That is because the difference between a decent email and good email is going to be barely something that matters. Or maybe you will use AI as a recreation tool, imagining and animating fantasy characters. Good fun but after a few hours or days you are going to be either bored, or would move on to focus on your actual work.
In contrast, the rich are increasingly using AI for tasks at which AI also uncannily excels. For example, our AI evangelist recently used Claude to find an airline stock to invest in. Claude ran an analysis, told him to discard what he wanted to buy and instead suggested another stock. He bought what Claude suggested. The AI proved right, within days he saw his investment grow significantly.
Similarly, one of the biggest groups of people who are finding the AI magical are software engineers and people in IT. These are also some of the richest people in the world and because their work involves dealing with definitive logic, exactly the kind of work that AI excels at, they are finding AI incredibly useful at coding work.
Essentially, whether you find AI magical or not depends greatly on what you are trying to do. And currently AI seems to be excelling at the kind of work that rich people do — that is consulting, research, dealing with data and financial analysis, coding, and planning. This is the reason why employees in a company might find that AI is leading to more work slop. But c-suite executives have a different experience. The work of the two groups is different, and at the moment AI is simply better at the kind of things that rich people do.
An alarm bell ringing?
This brings me to the final point of this piece — the alarming trend of AI not just highlighting the rich-poor divide but accentuating it. In a world where the best AI tools are behind steep paywalls, and in a world where best AI tools work to maximise capital instead of welfare or material productivity, chances are that the rich-poor divide will take on a shape that can never be bridged.
Already a number of concerns have been raised. AI godfather and Nobel Prize winner Jeffrey Hinton has categorically said that AI would increase the rich-poor divide at the global scale. In 2025, he said, “AI will make a few people much richer and most people poorer.”
Then a UN report earlier in December highlighted how AI was already creating a gulf between rich north and poor south. Do remember that currently in terms of using AI, largely because of how expensive it is and the kind of work it does, the rich countries rank far ahead of developing nations. No wonder the UN report said, “We think that AI is heralding a new era of rising inequality between countries, following years of convergence in the last 50 years.”
But the starkest warning, in my opinion, came around a decade ago from Yuval Noah Harari. Writing in his Homo Deus, a book about the perceived future of humanity, Harari wrote: “In the early twenty-first century the train of progress is again pulling out of the station – and this will probably be the last train ever to leave the station called Homo Sapiens. Those who miss this train will never get a second chance.”
When Harari wrote this in 2016 he did not have the name for his metaphorical train. Ten years later in 2026 the world has. It is called AI. And currently the ticket to this train is prohibitively expensive for the majority of people.
– Ends
Source link
[ad_3]
