The price you're shown isn't really about the product. It's about you.

There's a strange feeling that a lot of people have had without quite being able to explain it. You book a ride, and the fare looks higher than it did last week for the same trip. You're offered a shift, a rate, a deal, and something about it feels personal, even though you know it came from a machine. It turns out that feeling has a name, or rather two names, because two different researchers have been tracking the same phenomenon from opposite ends.

On the consumer side, it's called surveillance pricing. On the worker side, it's called algorithmic wage discrimination. Strip away the labels and they describe the same underlying mechanism. A company gathers behavioural data on you, whether that's your app usage, your location patterns, your spending habits or how quickly you say yes to things. It uses that data to sort you into a segment alongside people who behave similarly. Then it runs a live experiment on that segment to work out the maximum price you'll accept, or the minimum wage you'll tolerate. Not once. Continuously. Every interaction becomes another data point that nudges the number a little further in the company's favour.

This isn't hypothetical. It's already reshaping two industries that most of us interact with regularly, and the mechanics are worth sitting with because they show exactly why this matters as an ethical question, not just a commercial one.

The gig economy version

Anyone who has driven for a ride-share platform will recognise the twenty second decision window. An offer appears on the screen. Distance, time, traffic, a number. You have moments to accept or decline before it disappears and goes to someone else. It feels like a snap judgement, and it is, but the platform behind it is doing something much slower and much more calculated.

If you accept a lower offer, the pattern that emerges over time is that your next offer tends to be lower too. The algorithm has learned something about your threshold, and it adjusts accordingly. If you're pickier and hold out for better offers, they tend to arrive. But that patience is expensive, and for anyone relying on this as a genuine source of income rather than the odd extra shift, the pressure to accept eventually wins out. Once it does, the downward drift resumes. Whatever flexibility or leverage you started with erodes, quietly, one accepted offer at a time.

The unsettling part isn't that this happens. It's that it happens without a single conversation, without a manager ever deciding to cut anyone's pay, and without any of it being visible to the person on the receiving end. The system doesn't need to be cruel to produce a cruel outcome. It just needs to be relentless.

The nursing version

If the gig driving example feels like it belongs to a specific corner of the economy, the second example makes clear how far this logic travels. A number of newer staffing platforms in healthcare position themselves as a more efficient, tech-enabled alternative to the traditional local agencies that used to place contract and travel nurses. Some of these platforms, according to recent research from labour-focused think tanks, draw on data purchased from third party brokers, including signals about a worker's financial situation, to help determine the rate they're offered.

The logic is blunt once you see it stated plainly. Someone carrying more debt, or showing signs of financial strain, may be more likely to accept a lower rate simply because they need the shift more than someone in a stronger financial position does. The platform isn't assessing skill or experience to set that price. It's assessing desperation, and pricing accordingly.

What makes this worth pausing on is the shift in scale rather than the shift in intent. The impulse to pay people the least amount they'll tolerate is not new, and it is not unique to technology companies. What's new is the capacity to do it precisely, individually and continuously, across an entire workforce, without ever needing a room full of people making those judgement calls by hand. A task that would once have required enormous manual effort to run at scale now runs quietly in the background of an app.

Why this is an AI ethics question, not just a business one

It would be easy to read both of these examples as stories about bad actors doing bad things, and to conclude that the fix is simply better behaviour from a handful of companies. That's part of it, but it misses the more structural point. These systems aren't doing anything unusual by the standards of the technology itself. Segmenting audiences, running live experiments and optimising towards a target are standard machine learning practices, used constructively in countless other contexts. The technology is neutral. The question is what it's pointed at, and who it's accountable to when it gets there.

That's precisely why this belongs in a conversation about AI ethics rather than being written off as an inevitable feature of automation. The tools that make this kind of granular, continuous experimentation possible are the same tools that could be used to build fairer pricing, more transparent wage structures and systems that treat people as individuals rather than as segments to be tested against. The difference isn't the algorithm. It's the intent behind it, the transparency around how it operates, and whether the people affected by it have any visibility into what's happening to them.

We talk a lot, in this space, about AI as a force that can genuinely improve lives when it's guided by people who take that responsibility seriously. Surveillance pricing and algorithmic wage discrimination are useful precisely because they show the alternative so clearly. They're a reminder that the same capability can just as easily be turned against the people it touches, and that the difference between the two outcomes isn't technical. It's a choice, made by the people designing and deploying these systems, about who they're actually meant to serve.

If you've ever had that faint, hard to articulate feeling that a price or a wage was somehow calibrated to you specifically, it's worth taking seriously. Increasingly, it probably was.