Hostile Architecture: 10 of the Most Inhumane Examples, and Why They Don’t Work
There’s a strange thing that happens once someone explains hostile architecture to you. You walk out of the house, and suddenly you can’t stop seeing it.
This is design doing political work quietly. And because it’s quiet, it mostly goes unchallenged — which is precisely why it’s worth naming.
What is hostile architecture?
Hostile architecture is the use of physical design to discourage specific behaviours in public space. Sleeping. Sitting for too long. Skateboarding. Gathering. Lingering anywhere someone has decided lingering is unwelcome.
It goes by a small library of names depending on who’s talking. Councils and developers tend to say defensive architecture or defensive urban design, which sounds protective. Academics have used disciplinary architecture, exclusionary design and unpleasant design. Critics say anti-homeless architecture, because that’s who bears the weight of it. The vocabulary you choose is already an argument.
What all the terms describe is the same move: instead of policing a behaviour, you build an environment where the behaviour is physically impossible. No signage, no enforcement officer, no confrontation. The bench simply cannot be slept on, and nobody has to say a word.
Supporters argue this is ordinary stewardship — keeping shared spaces safe, clean and usable for everyone. Critics point out that “everyone” is doing an enormous amount of work in that sentence, and that the people quietly removed from the definition are the ones with nowhere else to go.
Here’s the part that gets missed most often: hostile architecture rarely inconveniences only its intended target. A bench you can’t lie on is also a bench a pregnant woman can’t stretch out on, an elderly man can’t rest properly on, and a disabled person can’t use comfortably. A plaza with nowhere to sit excludes the office worker on a break as efficiently as it excludes the rough sleeper. When you design against one group, you tend to design against a much wider slice of the public than you intended — and the wider public usually doesn’t notice, because they have somewhere else to be.
Why do cities use hostile architecture?
Three reasons come up repeatedly, and it’s worth taking them seriously rather than dismissing them.
It’s cheap and immediate. A set of spikes costs a few hundred pounds and can be installed in a morning. Housing costs millions and takes years. A council under pressure from local businesses about people sleeping in a doorway can act this week, visibly, for almost nothing. That’s an enormous political temptation.
It comes wrapped in a respectable theory. Much of this traces back to Crime Prevention Through Environmental Design and defensible space theory — a genuine body of urban design thinking about how environments influence behaviour. Some of that work is sound. But the framework has been stretched to cover measures its originators never had in mind, and “designing out crime” slides easily into designing out people.
Nobody has to take responsibility. This is the quiet advantage. A police officer moving someone on is a person making a decision, and can be argued with, filmed, complained about. A sloped bench is just a bench. There’s no one to appeal to. The exclusion has been outsourced to an object, and objects don’t have to justify themselves.
Common forms of hostile architecture
Spikes and studs
The most notorious form, and the one that reliably causes public outrage when photographed. Metal studs set into doorways, ledges, alcoves and paving — anywhere flat enough and sheltered enough that a person might lie down.
They became a global talking point in 2014, when studs appeared outside a London residential block and the images travelled fast. The reaction was swift enough that several retailers and property owners removed similar installations rather than defend them. That episode is instructive: hostile design tends to survive on obscurity, and rarely survives a photograph.
Benches you cannot lie on
The single most widespread category, and the most deniable. Central armrests, individual bucket seats, sharply sloped surfaces that make you brace with your legs, and perches designed for leaning rather than sitting.
Every one of these can be defended as a comfort feature or a maintenance decision. Some genuinely are. That ambiguity is the point — it’s what makes the bench such an effective instrument. You cannot prove intent from a photograph of an armrest.
Boulders, planters and bike racks
Objects placed not for their stated function but for the space they occupy. Landscaping boulders under flyovers. Planters that appear overnight on the exact patch of sheltered pavement someone was sleeping on. Bike racks installed where no cyclist has ever parked.
Portland is the frequently cited case, where boulders were placed at sites where roadside camps had formed. Because these installations look like amenities, they attract almost no scrutiny — which makes them the most quietly effective tactic on this list.
Fences and grates
Barriers restricting access to the places that offer shelter: under stairways, beneath bridges, over the warm-air vents that are among the few reliable heat sources on a winter street.
In 2015 the City of Stockholm installed a fence at a cost of around 200,000 Swedish krona to stop people sheltering beneath a stairway on Kungsholmen. Worth pausing on that figure, because it recurs across every case in this article: the money is real, it is available, and it is being spent on prevention rather than provision.
Sound, light and surveillance
The forms that don’t look like architecture at all. High-frequency deterrent devices audible mainly to teenagers. Classical music piped through transport hubs to discourage loitering. Blue-tinted lighting in public toilets, intended to make veins harder to find. Cameras, which don’t physically prevent anything but shift behaviour through the expectation of being watched.
Surveillance sits slightly apart from the rest — it’s a different mechanism, working on the mind rather than the body. But it belongs in the same conversation, because it’s aimed at the same outcome: making a space feel like it isn’t yours.
10 hostile architecture examples from around the world
1. Concrete spikes under the expressway, Guangzhou, China
Roughly 200 square metres of concrete spikes beneath the Huangshi expressway, making the ground physically unusable. Residents reported that people had previously gathered and socialised in the space beneath the bridge — so the installation didn’t only remove shelter, it removed a small piece of neighbourhood life along with it.
2. Anti-loitering spikes on a bank façade, Mumbai, India
Rows of sharp metal spikes ringing the base of an HDFC Bank branch. The bank described them as an anti-loitering measure. Local residents raised an obvious and uncomfortable point: spikes at that height don’t distinguish between someone choosing to sit and a child, an older person or a disabled pedestrian losing their balance.
3. Curved and sloped benches
You’ve almost certainly sat on one and quietly wondered who signed it off. A convex or angled seat is very difficult to lie on — gravity does the enforcement for you. It’s also, for anyone with a bad back or hip, a genuinely unpleasant place to sit for the twenty minutes a delayed bus requires.
4. The Camden bench, London, England
The Camden bench has been called the ultimate anti-object, and it earns the title. Introduced in 2012, it’s a solid slab of concrete engineered against a checklist of unwanted uses at once: no flat surface to sleep on, no grindable edge for skateboards, no recess to stash anything in, no gap for litter, nothing to steal, nothing easy to graffiti.
What makes it the most revealing object on this list is how much design effort went into it. Someone sat down and worked hard, iteratively, thoughtfully, to produce a bench optimised against people. Whatever else you think of it, that is a serious piece of design — pointed in a direction worth arguing about.
5. Barred and blocked corners
Even a recessed corner — a sheltered metre of pavement out of the wind — gets caged off. Officially the aim is discouraging panhandling and loitering. Practically, it removes one of the last free, unsupervised places a person can stand out of the rain.
6. Armrests as architecture
Armrests genuinely help many people get up from a seat, and that’s a real accessibility benefit. But look at the placement. An armrest at each end is furniture. An armrest every 50 centimetres along a three-metre bench is a policy.
7. Tube benches, Tokyo, Japan
Tokyo’s tubular benches are subtler than spikes and arguably more effective. The rounded profile makes lying down impossible, and the bare metal handles the rest — scorching in summer, freezing in winter. No one is excluded by name. The material simply makes staying unpleasant for everybody, which is a strategy in itself.
8. Planters as street dividers
Greenery in the city is almost always good news, which is exactly what makes this tactic so hard to challenge. Planters installed along a stretch of sheltered pavement look like an improvement. They also happen to eliminate the flat, dry ground someone was sleeping on.
Student newspaper The Eye Opener documented one such case in Toronto, reporting that a man who regularly slept outside a coffee shop on Victoria and Dundas was asked by police to move across the street so planters could be installed — to a side of the road with no shelter from rain or wind. That detail is the whole subject in miniature: the problem wasn’t solved, it was relocated somewhere worse.
9. A public plaza with nowhere to sit, Manhattan, New York
A large public plaza at East 56th Street and Third Avenue, with no tables and no chairs. Office workers on a break end up leaning against a wall. One worker quoted by The New York Times summed up the message the space sends: it doesn’t feel public, and it feels like people are being kept from using it.
This is hostile architecture at its most sophisticated, because there’s nothing to photograph. No spike, no bar, no bolt. The hostility is in what was never installed — and absence is very difficult to protest.
10. Fenced-off heating grates
In a cold city, a warm-air grate is one of very few free heat sources available to someone sleeping outside. Fencing them off is among the starkest measures on this list, because the thing being denied isn’t comfort or dignity but warmth, in weather that can kill.
Does hostile architecture actually reduce crime or homelessness?
The honest answer is that there’s very little evidence it does either — and a fair amount suggesting it makes things worse.
The core problem is that hostile design doesn’t touch a single cause of homelessness. Nobody has ever become housed because a bench got an armrest. What changes is location: people move to somewhere less visible, which in practice means somewhere less safe, further from services, and further from the passing witnesses who make a public street safer to sleep on than an alley.
Researchers looking at this have been fairly consistent. The designs are widely described as cruel and ineffective, displacing and worsening the crisis rather than easing it, while pushing unhoused people into more isolated and dangerous places. There’s a second cost too, harder to measure: exclusionary design damages social trust and degrades the health of public space for everybody who uses it.
The crime argument fares no better. Deterring a behaviour in one location isn’t the same as preventing it, and displaced activity has a habit of reappearing a street away. What the measures reliably produce is a specific visual outcome — a plaza that looks orderly — which may be the actual objective, even when it isn’t the stated one.
Is hostile architecture legal?
Almost everywhere, yes. And in the United States the legal ground moved sharply in 2024.
In City of Grants Pass v. Johnson, decided on 28 June 2024, the US Supreme Court ruled 6–3 that enforcing ordinances banning camping on public property against homeless people does not violate the Eighth Amendment’s prohibition on cruel and unusual punishment. The decision overturned the Ninth Circuit’s earlier position in Martin v. Boise, which had held that cities couldn’t punish people for sleeping outside when no shelter was available.
The practical consequence is that cities can now enforce camping bans through fines and jail time even where there is nowhere for people to go. The effects came quickly — California, home to roughly a third of the country’s unhoused population, saw the governor order state agencies to clear encampments within about a month of the ruling.
Why this matters for design: hostile architecture and criminalisation are two halves of the same approach. One makes the space unusable, the other makes the use punishable. Where the law has hardened, the built environment usually follows.
Elsewhere the picture is more mixed. Some UK councils have quietly removed installations after public pressure, and a growing academic literature — including work published in 2025 arguing that anti-homeless hostile design constitutes wrongful discrimination — is starting to build a case that current law doesn’t yet reflect.
Who is fighting back against hostile design?
The most effective resistance so far hasn’t come from planning committees. It’s come from people with cameras.
British artist Stuart Semple photographed metal bars retrofitted to benches in his hometown of Bournemouth in 2018 and posted the image. It was viewed more than a million times within 24 hours. He then organised a public intervention inviting residents to decorate the benches and turn them into “love benches” — and that night, council contractors quietly removed the bars in the dark.
That episode became the basis of HostileDesign.org, which distributes “design crime” stickers so anyone can tag hostile installations in their own city, and aggregates photographs submitted from around the world. Semple’s definition is broader than homelessness — he describes hostile design as design intended to restrict freedom or control people, whether they’re homeless, skateboarding, or simply gathering somewhere to enjoy themselves.
In 2024 he went further, working with agency TBWA\MCR on a London campaign that fitted life-size photographs of sleeping rough sleepers over real doorway spikes, so the actual spikes pierced the images. It’s a brutal piece of communication, and it works for the same reason the Bournemouth photo did: it makes visible something engineered to be overlooked.
The pattern across all of these is consistent, and it’s the most useful thing in this article for anyone who wants to act. Hostile design almost never survives sustained public attention. It depends on nobody looking closely.
What are the alternatives to hostile architecture?
This is where the conversation usually stalls, because “just build housing” sounds like a non-answer to a council with a bench-shaped problem this month. So it’s worth being specific.
At the level of street furniture: benches without central bars, seating with backs and armrests at the ends only, shaded and sheltered rest points, public toilets that are actually open, and drinking water. These serve older people, disabled people, parents, workers and tourists — and they happen to also serve people with nowhere to go. Inclusive design isn’t a compromise; it’s better design for a wider public.
At the level of policy: the strongest evidence points at Housing First, which inverts the traditional sequence. Instead of requiring sobriety or treatment before housing, it provides a tenancy first and support afterwards. Finland built national policy around it from 2008, and single homelessness fell by roughly 45% over the following decade, making it the only European country where homelessness consistently declined. Finland’s environment ministry has estimated savings of up to €15,000 a year per formerly homeless person through reduced emergency healthcare, social services and justice costs.
It would be dishonest to present that as a solved story. Finland’s numbers have reversed sharply since — single homelessness rose about 11% in 2024 and roughly 20% in 2025, the largest annual increases in its records, following policy changes and funding pressure. The lesson isn’t that Housing First failed. It’s that it worked while it was funded and faltered when it wasn’t, which tells you something fairly precise about what actually drives the numbers.
Either way, the comparison stands: one approach measurably reduced homelessness, the other measurably relocated it.
What can architects and designers actually do?
Someone drew every object in this article. Someone specified it, someone approved the budget, someone signed the installation off. Hostile architecture isn’t weather — it’s a series of professional decisions, which means professionals can decline to make them.
Three practical positions worth holding. Ask what a feature is for. A central armrest has a stated purpose and an actual purpose, and the gap between them is where this whole subject lives. Ask who a design excludes. Every design excludes someone; the question is whether you’ve noticed and whether you can defend it. Ask whether you’d defend it publicly. The Bournemouth bars disappeared overnight once photographed, which suggests the people who installed them already knew the answer.
None of this is new to the profession, either. The question of who buildings serve and who they quietly turn away runs right through the history of the discipline — a thread you can follow through the work of the most influential architects in history, whose reputations rest as much on who they designed for as on what they built. It’s also increasingly taught rather than assumed; design ethics and inclusive practice now feature seriously in the curricula at many of the leading architecture schools, which is a genuinely encouraging shift from twenty years ago.
For a broader view of where design energy is currently being directed, it’s worth reading about 5 hot technologies transforming contemporary architecture and how visions of futuristic architecture have changed over 100 years — both a reminder that the profession’s imagination is capable of far more ambitious problems than keeping people off a bench.
The uncomfortable conclusion
Hostile architecture works. That’s the part critics sometimes skip past, and it’s worth conceding, because it explains why the practice persists.
It works at the thing it’s actually designed to do, which is not reducing homelessness but reducing the visibility of homelessness. The doorway is clear. The plaza looks tidy. The complaints from the shop downstairs stop. Measured against that brief, the spikes are a complete success.
The trouble is that it’s a brief no one will state out loud, because stated plainly it’s indefensible: we would like these people to be somewhere we can’t see them, and we’re prepared to spend public money on hardware to achieve it. Every justification about safety and order is a way of avoiding that sentence.
Nobody has ever been housed by a bench. But a bench can tell you, with total precision, what a city thinks about the people who need one.
Tags: Architecture and SocietyHostile ArchitectureHUB — Urbanism- Landscape & SustainabilityPublic SpaceUrban DesignUrban Planning
Emily Reyes is a Brooklyn-based architecture writer and Article Curator at Arch2O, known for her sharp eye for experimental design and critical theory. A graduate of the Southern California Institute of Architecture (SCI-Arc), Emily’s early work explored speculative urbanism and the boundaries between digital form and physical space. After a few years in Los Angeles working with boutique studios on concept-driven installations, she pivoted toward editorial work, drawn by the need to contextualize and critique the fast-evolving architectural discourse. At Arch2O, she curates articles that dissect emerging technologies, post-anthropocentric design, and contemporary spatial politics. Emily also lectures occasionally and contributes essays to independent design journals across North America.


















