With the arrival of AI, everyone has access to all the knowledge that is available online, without needing to be able to find or analyse it themselves. It is all done for you. You have a lawyer, a research assistant, a marketeer, a programmer and a therapist in your hands, or a combination of all of them, and it never sleeps.
Before, access to all of this was gated. It cost money, or it cost a lot of time to learn the skills yourself. Now it is commoditised.
If you belonged to the group that had difficult access to knowledge or professional help, because of a lack of financial resources, then for you this is a revolution that will change your life, if you become aware of it.
So, yes: AI is also making the world more inclusive in some sense. Let’s dive deeper into that, before we move to the negativity.
A door that opens
We should celebrate the positive effects of AI, the same way that we should condemn the negative ones.
People who have been excluded before can now take part. Think about all the people who have difficulty writing correctly in English, or even reading. In the past, you could get disqualified if your writing level wasn’t great. You were excluded, without anyone really looking at what you had to say.
And yes, until now we have seen AI-written text as lazy. But wait until it all gets accepted, or even becomes unrecognisable. Then suddenly these people’s arguments and offers are valued for their content, and not assessed by their form.
Also, someone who could never hire a designer in the past can now still have decently designed content. Or, to go even further, a person could even produce a video to express his feelings about a subject, which is the ultimate level of freedom of expression.
Apply this to business: a single person, who could never pay a marketing agency before, is now able to create a logo, a full brand identity, and a website in an afternoon. And he can make campaign material that stands next to what a funded competitor puts out.
For people who normally wouldn’t get an opportunity, AI gives them an entrance to the market. It gives them access, and a real chance to matter in a market that was closed to them before.
And this is not just a feeling. Researchers at Stanford and MIT studied more than five thousand customer-support workers and found that the least experienced and lowest-skilled people gained the most from AI, and even improved their written English along the way. The people who were furthest behind moved forward the fastest.
This is something to celebrate. The market is becoming inclusive, and even more so with the progress being made in generative AI.
But being allowed in is not the same as winning
Almost everyone has better access now. But if we look at the other benefits, beyond inclusion and access, a different story can be told. Inclusive does not automatically mean fair, and we keep confusing those two things.
The middle class is the best example. People here are losing some of their privileges. They have been trained to sell their hours as professionals, with a very narrow expertise they have built up over the years. With the arrival of generative AI, they will see this become less and less interesting for buyers over time.
The same AI that hands someone a (close to) free developer and design studio is also putting a lot of jobs under pressure. And often it is the same kind of people who feel both sides of it.
Knowledge work used to be the way up. Think about a paralegal, a junior analyst, a translator, a support agent or a copywriter. These were the first steps you could take to climb. And these are exactly the steps that AI is taking over right now. And a bit later, the higher-paid jobs will follow.
And this is not a prediction anymore. A 2025 Stanford study found that young workers between 22 and 25, in exactly these kinds of AI-exposed jobs like software development, accounting and customer service, saw their employment fall by around 13% since the end of 2022. Older workers in the same roles barely moved.
Then there is something else happening. While almost everyone can use AI now, very few people can actually shape it. A new elite is forming around the ones who understand the technology deeply, who build the models and own the infrastructure.
And they are the ones who get the power and the financial benefits.
The new AI-elite benefits the most
Here we have to separate two things: who can use AI, and who it is actually made for. The first is obvious. The latter is about who pays for AI, and who it is designed to pay and ultimately serve. If we go deeper and follow the money to answer these questions, we get to a painful truth: it only serves a small group of people.
Let’s look at who is directly becoming rich from AI first. It’s a very small group that is on the receiving end of the direct financial streams created by the AI wave (or bubble). They build and sell the AI itself, and they get paid every single time the rest of us use it. The one-person business that now builds its own brand, without a designer and without an agency, is still paying rent. Quietly, every month, to the handful of companies that own the models and the chips underneath them. We escaped the old gatekeepers and walked straight into a new one. Only this time it is almost invisible, folded into a subscription, and a very, very small group is becoming extraordinarily rich by providing it.
And it’s not the only way we’re paying for it. A big part of the most advanced AI is funded by military budgets, and in the end those budgets are paid by us too, as taxpayers. Look at all the wars that are happening as I write this. The budgets are almost unlimited.
The people who die in these wars are mostly poor. It does not help humanity, it only helps the tech giants and the armies with the biggest budgets.
Now, on the consumer side, AI has been leveraged for years to make platforms as addictive as possible, simply because attention is what gets sold. Meta sells ads, so they want the user to see as many as possible. That’s only possible if you stay on the app longer.
Same goes for TikTok, or any other social platform.
The ones who lose the most are the people who have the least space to deal with it. And so the poorest end up paying for a “free” product with their time and their attention, which they never get back.
Then there is the use of AI in the stock and crypto markets, which makes people who are already rich even richer, and now it happens faster.
And finally there is everything built to keep people out. Fraud detection, loan approvals, insurance pricing, hiring filters. All of these decide who gets in and who gets refused. And again and again, these systems discriminate against the same groups that were always excluded, because they are built to learn from the old behaviour and to scale it… The only difference is that a non-transparent layer is added to the process, and now it looks neutral and fair.
And this is not hypothetical. A widely used American health algorithm, studied in the journal Science, treated Black patients as healthier than they really were, only because less money had been spent on their care in the past. Amazon had to scrap a hiring tool after it taught itself to mark down CVs from women, because it had learned from years of mostly male hires. And research on American mortgages found that even when the loan is approved, Black and Latino borrowers are charged more, hundreds of millions extra every year. A higher price is still a barrier. And for someone with little to spare, that barrier is its own quiet form of exclusion.
More inclusive, but not more equal
So, back to the question: is AI making the world more inclusive? Yes, it is. But it does not benefit everyone equally, and that is the part we cannot ignore.
AI gives knowledge, a voice and a design studio to people who never had any of them. And at the same time, its strongest incentives pull it towards the people who already have both: the power and the money.
The thing I keep thinking about is this: the inclusive side mostly happens by accident. It comes almost for free, because a general tool is cheap and available to everyone. The other side, the part that serves power, is built on purpose. Because that is where the budgets are, and that is where the returns are.
That accidental, inclusive side is exactly the part we should magnify, to tip the balance in favour of inclusivity. The outcome is not fixed yet. It depends on what we build, who we build it for, and why.
So if you are building something with AI, keep this in mind: everything you build serves someone. Be critical about who you serve.
Try to build the door that lets people in, instead of the filter that keeps them out. Or, even worse, kills them.
