Dehumanising technology almost never looks like a villain. There's no glowing red eye, no cackling. It looks like a faster checkout, a tidier spreadsheet, a chatbot that answers at 3am. The harm slips in sideways, dressed as progress — which is exactly why it's worth learning to see.

To "dehumanise" simply means to treat a person as something less than a full human being: as a number, a data point, a problem to be processed, or a source of attention to be mined. Technology doesn't have to do this. But when it's designed only for efficiency, profit or scale — with people treated as inputs rather than the point — it drifts there on its own.

Here are six of the most common patterns. None of them require bad people to occur. They mostly require no one stopping to ask, how will this feel from the other side of the screen?

1. You get reduced to a metric

The first move is always the same: a whole person is flattened into a single measurable thing. A delivery driver becomes a "drop rate". A patient becomes a bed number. A reader becomes "time on page". A child becomes a reading level.

Measurement isn't the enemy — we need numbers to run anything at scale. The danger is when the metric quietly replaces the person in everyone's mind, including their own. We start optimising the number instead of caring for the human it was meant to describe. The map eats the territory.

The moment a number stands in for a person, someone stops being seen.

2. The human exit is removed

You have a problem that doesn't fit the menu. You press 0. You type "speak to a person". You search the site for a phone number that no longer exists. The system was built so that, past a certain point, no human can be reached — because humans are expensive and "deflection" looks great on a dashboard.

This is one of the most widely felt forms of dehumanisation, because it removes the thing that makes us social animals: the ability to be heard by someone who can understand and act. Being stuck in a loop with no human at the end isn't just annoying. It quietly tells you that your particular situation doesn't matter enough to warrant a person.

3. Decisions are made about you that no one can explain

An application is declined. An account is frozen. A post is removed. A price doubles. When you ask why, the answer is some version of "the system decided" — and no one, not even the staff in front of you, can tell you how or appeal it in plain terms.

Opaque automated decisions strip away two deeply human things at once: understanding and recourse. We can usually live with a "no" if we know why and can respond to it. What corrodes dignity is a verdict from a black box, delivered as if it were the weather.

4. Your attention gets farmed

Some products are designed to help you do a thing and leave. Others are designed to keep you there — because their business depends on the hours you spend, not the value you get. Infinite scroll, autoplay, variable-reward notifications, streaks, and feeds tuned to provoke rather than inform are not accidents. They're the engineering of compulsion.

The dehumanising part isn't that these tools are fun. It's that they treat your attention as a resource to be extracted rather than a finite, precious part of your one life. You came to connect or relax; you leave foggy, agitated, and somehow worse about yourself — and the design counts that as a win.

5. You're watched, scored and sorted without consent

Surveillance has become ambient. Location trails, purchase histories, keystroke timing, productivity scores, "sentiment" read from your messages — gathered continuously, often invisibly, and used to predict and nudge what you'll do next.

Being constantly observed changes us. We perform instead of live. We self-censor. The sense that you can be yourself in private — a basic condition for being a person rather than a subject — gets thinner. And when that data is used to sort people into categories that shape what jobs, prices or opportunities they see, surveillance turns into quiet, automated discrimination.

6. The very language treats people as raw material

Listen to how technology talks about the people it serves. "Users." "Eyeballs." "Traffic." "Churn." "Engagement." "Monetising the audience." "Capturing" attention. None of these words contain a human face. Language shapes thought, and an industry that calls people "users" for long enough starts to design for use rather than for flourishing.

This one matters because it's upstream of all the others. Change how a team speaks about the people on the other side of the screen, and you change what feels normal to build.

Why naming it is the first step

Dehumanisation thrives on seeming inevitable — "that's just how the internet works now". Once you can name the pattern, it stops being weather and becomes a design decision someone made, which means a different decision is possible. You can ask for the human exit. You can choose tools built to respect you. If you build technology, you can refuse to ship the dark pattern even when the dashboard would reward it.

That's the whole reason this resource exists. Seeing clearly is not pessimism. It's the start of building something better.

Whether you write the code, sign the purchase order, or decide how a tool gets rolled out across a team, you have more power to keep technology human than you might think. Most dehumanising design isn't malicious — it's the result of no one being responsible for the human experience. These principles give that responsibility a home.

1. Keep a human in the loop — and a human at the door

Automate the repetitive, the predictable, the boring. But for anything that meaningfully affects a person's money, health, livelihood, reputation or access, make sure a human can step in, review and override. Just as importantly, make the human reachable: a clearly signposted way to reach a real person, without a maze, is the single most humane feature most services lack.

The test: if this goes wrong for one person in a way the system didn't anticipate, what happens to them? If the honest answer is "nothing — they're stuck", you have more design to do.

2. Make the person the customer, not the product

Be honest about who the technology actually serves. If the people using it are also the ones paying for it — directly or through their organisation — their interests and the product's interests can align. If they're "free", their attention or data is usually the thing being sold, and the design will drift toward extracting it.

You can't always change the business model, but you can refuse to let it quietly override the user's interests. When the two conflict, decide in advance whose side the product is on, and design accordingly.

3. Default to dignity

Defaults are the most powerful design decision you'll make, because most people never change them. So set them where you'd want them set for someone you love: privacy on, data collection minimal, notifications calm, the gentlest option first. Make the respectful choice the easy one and the extractive choice the deliberate one — never the reverse.

The default is your real value statement. Everything else is marketing.

4. Be explainable and appealable

If your system makes or shapes decisions about people, those people deserve to understand them and to push back. That means a reason a normal person can follow ("declined because X"), not a reference number. It means a real route to appeal that reaches a human with the power to change the outcome. And it means logging enough that you yourself could reconstruct why the system did what it did.

This is doubly true for anything using AI. A model's confidence is not the same as being right, and "the algorithm decided" is never an acceptable final answer to a person whose life it touched.

5. Design for the worst day, not the demo

Products are designed for the cheerful average case and demoed to applause. But people meet technology on their worst days too — grieving, panicking, locked out, in pain, in a second language, on an old phone with one bar of signal. Humane design asks: what happens to the most vulnerable person who will ever touch this?

If your error messages assume calm, your flows assume confidence, and your accessibility is an afterthought, you've designed only for the people who least needed your care.

6. Measure what's human, not just what's easy

We optimise what we measure, so be careful what you measure. Engagement, time-on-app and clicks are easy to count and dangerous to chase — they reward compulsion as readily as value. Sit a human metric next to every efficiency one: Did people accomplish what they came for? Did they leave better off? Would they recommend it to a friend they care about? Did we reduce, not just shift, someone's stress?

A genuinely good product is often one people use less over time because it actually solved their problem. Make sure your dashboards can tell that story without punishing it.

7. Move at the speed of trust

When adopting technology in an organisation — new software, automation, AI — the temptation is to deploy fast and explain later. That's how you get fear, quiet sabotage and tools that technically work but humanly fail. Bring the people affected in early. Tell them honestly what changes and what doesn't. Show how it makes their work better, not just cheaper. Train properly. Leave room to say "this isn't working".

Technology adopted with people lands; technology done to people gets resented, and resentment is its own kind of dehumanisation.

Questions to ask any vendor

If you're buying technology rather than building it, your leverage is in the questions you ask before you sign:

• Where does our people's data go, who can see it, and how do we get it deleted?
• If your system makes a decision we disagree with, how do we override it?
• What does support actually look like for our end users on a bad day?
• What are you optimising for, and how would we know if it started working against our people?
• If we leave you in two years, how do we get our data and our people's work out?

A vendor who welcomes these questions is one you can work with. A vendor who can't answer them is telling you something important.

The encouraging truth

None of this asks you to be slower, poorer or less ambitious. Humane technology is usually better technology: more trusted, more loyal, less likely to blow up into a scandal, more pleasant to build and to work for. Respect for people isn't a tax on good engineering. Increasingly, it is good engineering. The teams that figure that out first won't just feel better about their work — they'll win.

First, the most important thing: if technology has been making you feel small, anxious, invisible or exhausted, that's a reasonable response to environments often designed to produce exactly those feelings. The fault is not in you for struggling. And there's a lot you can do — none of it requiring you to throw your phone in a lake.

Pick one or two of these to try. You don't need all of them, and you don't need to do it perfectly. Small recoveries count.

Reclaim your attention

Your attention is being competed for by teams of clever people. You don't have to out-discipline them — you just have to change the defaults so willpower isn't doing all the work.

• Turn off most notifications. Keep alerts from actual humans (messages, calls). Silence the rest — the badges, the "someone you may know", the nudges to come back. You decide when to check, not the app.
• Make the addictive apps duller and further away. Move them off your home screen, into a folder, or log out so opening them takes a deliberate step. Greyscale mode makes a surprising difference.
• Put a small gap between impulse and action. A two-tap login, a screen-time reminder, or simply leaving the phone in another room while you eat or sleep returns the choice to you.
• Follow fewer things, on purpose. Curate feeds toward what leaves you informed or glad, and unfollow what leaves you agitated. You're allowed to make your inputs kinder.

Insist on reaching a human

When a service traps you in a loop, the goal is to get to a person with the power to help. A few moves that genuinely work:

• In a phone menu, try pressing 0 repeatedly, or saying "agent" / "representative" / "complaint". Saying you want to cancel often routes you to a real, empowered human fast.
• With a chatbot, type plainly: "I need to speak to a person." Repeat it. Don't argue with the bot — just keep asking for the handoff.
• Write down names, times and reference numbers as you go. It steadies you and makes any follow-up far stronger.
• If you're stuck, escalate sideways: a public message to the company, a formal complaint, or your country's relevant ombudsman or disputes scheme. Knowing the escalation path exists takes away the trapped feeling.
• Be firm and be kind. The person who finally answers usually didn't design the maze — they're stuck in it with you.

You are allowed to take up space, ask for a human, and expect to be treated like one.

Set boundaries you can actually keep

Boundaries fail when they rely on heroic self-control. Build ones the environment helps you keep:

• Pick tech-free zones or times — the dinner table, the first and last 30 minutes of the day, the bedroom. Charge the phone outside the bedroom and use a normal alarm clock.
• Batch the draining stuff. Check email or messages at set times rather than continuously. The world copes; you recover hours.
• Separate work from you. If you can, keep work apps off your personal phone, or use a focus profile that switches them off after hours. You are not a server that must always be available.
• Let people know. "I check messages a couple of times a day" resets expectations and gives others permission to do the same.

Choose tools that treat you like a person

You have more choice than the loudest apps suggest. Where you can, vote with your feet:

• Prefer tools you pay for over "free" ones when the free version is clearly selling your attention or data — paying is often the cheaper deal in the end.
• Favour products that are calm by design: no endless feed, no manipulative streaks, easy to leave.
• Do a quick privacy tidy-up: review app permissions and turn off location, microphone and contact access for anything that doesn't truly need them. Skim the privacy settings of your main accounts once and dial them down.
• When something respects you, tell people. Recommending humane tools is how they win.

When it's your employer's system

Sometimes the dehumanising tech is one you can't simply quit — a monitoring tool, a scheduling app, a rigid workflow. You still have moves:

• Name the specific harm, calmly and concretely. "The new system pings me on days off and I can't switch it off" is harder to dismiss than "I hate this app".
• Find your allies. If it's grinding you down, it's likely affecting colleagues too. Raised together — or through a union or staff forum where you have one — concerns carry far more weight.
• Know your rights. Many places have real legal limits on workplace surveillance, data use and the right to disconnect. A quick check of your local rules can be quietly empowering.
• Protect your energy where you can't change the system. Do the parts you control well, and don't let a badly designed tool convince you that you are the problem.

A gentle reminder to end on

You don't have to win against the whole machine. You just have to reclaim enough room to feel like yourself again — a quiet morning, a real conversation, an afternoon where your attention belonged to you. Each small boundary is a vote for a more human relationship with technology, and those votes add up, in your own life and in the wider culture.

Be patient with yourself. You're navigating environments built by experts to be hard to put down. Every bit of ground you take back is a real win — and it's yours to keep.