Co-op’s pavement robots revive an old argument about who the street is for

On an unremarkable afternoon in Milton Keynes, a small white box on six wheels noses its way along the edge of the pavement, keeping a cautious distance from a pram, pausing at the lip of a dropped kerb, then continuing as if following rules only it can see. It carries a bag of groceries ordered through an app, part convenience, part novelty, part experiment in whether the most ordinary acts of commerce can be automated without disturbing the fragile etiquette of shared public space.

This is the point at which the debate stops being about technology and becomes about the street itself. The Co-op’s robot deliveries, run in partnership with Starship Technologies, have existed in pockets of Britain since before the pandemic. Yet their growth now meets a sharper public mood. Living Streets, the charity founded in 1929 as the Pedestrians’ Association, has written to the Co-op’s interim chief executive, Kate Allum, warning that the machines “present a danger to pedestrians”. It is less a complaint about gadgets than a defence of pavements as civic infrastructure, not a warehouse aisle.

At first glance, delivery robots look harmless. Their speed is modest. Their size is manageable. Their design cues are deliberately friendly, closer to a footstool than a motor vehicle, inoffensive in a way that is clearly intentional. But that friendliness can obscure what Living Streets is really arguing: that each new commercial use of the pavement, however small, competes with those for whom the pavement is not an optional extra but the only viable route.

“The pavement is not a logistics network; it is a public space, won through decades of accessibility campaigning and relied upon every day by people for whom it is the only viable route,” Catherine Woodhead, Living Streets’ chief executive, said in the letter. It is an unusually forceful line, and it has resonated because it captures a national frustration that has been building for years. The modern pavement is already squeezed by parked cars with two wheels on the kerb, by street clutter, by pavement cafes, by e-bikes and e-scooters, by bins, by signage, by roadworks that last for months. The question is not whether a single robot can pass a pedestrian safely; it is whether the cumulative effect of turning pavements into commercial corridors makes everyday mobility harder, especially for those who already move through the world with greater difficulty.

Woodhead has emphasised older people in particular, and her framing is careful: for many, a short walk to buy a pint of milk may be their only chance to see other people. That detail matters because it casts the pavement as social infrastructure, not simply transport space. A robot that obliges someone to step aside, to pause, to negotiate, or to feel uncertain about what will happen next is not merely an inconvenience. It is a tiny tax on confidence, repeated often enough to change behaviour. When a charity talks about “accessibility”, it is usually describing the unseen margin between independence and isolation.

The Co-op’s role is also more complicated than the headlines allow. The retailer offers deliveries from a number of stores in places including Leeds, Milton Keynes and Northampton, but it does not take the robot orders itself. Customers order through Starship Technologies’ app. In practice that means a household may think it is using a Co-op service, while the operational decisions about routing, speed, stopping behaviour and customer interaction are largely those of a technology supplier whose business model depends on scale. That distinction will become more important as scrutiny grows, because accountability tends to follow the money and the brand, not the fine print of which app was used.

Starship is not a fringe start-up. Founded in Estonia in 2014 by Janus Friis and Ahti Heinla, the co-founders of Skype, it describes itself as the world’s largest self-driving robot manufacturer. Its machines have become familiar in certain test-bed towns, an emblem of the broader shift in how services once performed by people are being reorganised around software, sensors and remote supervision. Starship has also partnered with restaurant delivery firms such as Just Eat, signalling ambitions beyond groceries. What begins as a practical solution for short-distance deliveries can become, if successful, an entirely new layer of street traffic.

That prospect is what sits behind the regulatory pressure now gathering in Whitehall. Campaigners describe the situation as a “legal vacuum”. In reality, the machines exist in a grey zone, tolerated because they are small and because there is no clear category into which they fit. Yet ministers are reportedly considering adapting the laws around these robots so that they would be classified alongside e-scooters and other micro-mobility vehicles. That is a striking move, because it would acknowledge that delivery robots are not merely unusual pushchairs; they are vehicles whose presence needs explicit rules, duties and enforcement.

Britain has been here before with newer forms of movement. E-scooters arrived as a mixture of pilot schemes and informal adoption, with regulation chasing reality. The result has been patchy enforcement and a lingering sense that public space is being rewritten by trial and error. To place delivery robots in the same family as micro-mobility implies something similar: a recognition that innovation will proceed whether or not legislation is ready, and that the state must decide how much risk, clutter and inconvenience it is willing to allow while the market experiments.

Supporters of the robots will point out, correctly, that delivery itself has a footprint. Vans stopping on double yellows, drivers carrying crates, engines idling, repeated trips for small orders, all have costs in congestion, emissions and safety. If a robot can replace some of those journeys, there is a public interest in exploring the trade-off. There is also the consumer’s interest. A grocery delivery at short notice, with a small basket, is exactly the kind of task that turns uneconomic when labour is scarce or expensive. The robot promises to make that marginal service viable. In a sector where margins are thin and expectations of convenience are high, the attraction is obvious.

Starship has also sought to quantify benefits in ways that anticipate government priorities. A report commissioned by the company last year suggested that increasing the use of robot deliveries across Britain could boost household spending on groceries by about £125 million and lift productivity by saving the equivalent of about 300,000 working hours. Even allowing for the self-interested nature of commissioned research, the pitch is well targeted: growth, productivity, consumer spending. It invites policymakers to see a new industry rather than a new nuisance.

But these headline gains sit uneasily with the very ordinary, very local reality of pavements. Productivity saved for whom, and at whose expense? If the time saved is mainly that of customers who can already access delivery services, while the cost is borne by pedestrians who have no alternative route, then the distribution of benefit and burden is not neutral. Public space is one of the few assets shared by everyone, and its redesign is rarely experienced equally.

The international examples cited by Living Streets underline that this is not simply a British nervousness about new machines. Robots have been banned in some cities, including Toronto and parts of Chicago, amid reports of collisions with pedestrians and injuries. Such decisions usually emerge not from one dramatic incident but from a steady accumulation of complaints, near misses and uncertainty about liability. The precise details differ by city, but the pattern is consistent: once machines become common enough to be noticed, they become common enough to be regulated, and the pivot from novelty to nuisance can be swift.

Starship, for its part, has previously said it plans to build an entire new fleet in Britain if the government clarifies its regulatory stance, and has claimed that doing so would encourage wider investment in the sector. This is the bargaining position of a company that wants rules, but wants them shaped to make expansion easy. Clarity can protect the public, but it can also legitimise growth. The state is being asked, in effect, to choose between two kinds of uncertainty: the uncertainty of allowing machines to roam the pavements under loosely applied general principles, and the certainty that clear rules will accelerate adoption.

There is also the question of what “safety” means in this context. A robot may be less likely than a van to cause severe injury in a collision, yet still be capable of harm. More commonly, safety includes the risk of trips and falls, the panic of a sudden obstruction, or the distress caused to someone with impaired vision who cannot easily interpret the robot’s intentions. Humans are adept at reading human movement: eye contact, a slight turn of the shoulder, a wave. Machines must replace that with signals, lights, pauses, perhaps sounds, all of which can be misread. A pavement is not a controlled environment; it is a messy, social, unpredictable space.

Advocates of pedestrian rights are not simply resisting change. They are responding to a deeper issue: the repeated tendency for new commercial systems to treat the public realm as a resource to be exploited, rather than a shared asset to be protected. When Woodhead says the pavement is not a logistics network, she is pointing to a cultural shift in which delivery has become an expectation attached to almost anything we buy. The warehouse, the app, the algorithm, the last metre of the journey, all are designed for efficiency. The pavement, by contrast, was designed for people at walking pace, people who stop, hesitate, chat, or move slowly. The two logics collide.

The Co-op’s predicament is that it is both a participant in this transformation and a retailer with a brand that trades on community. Convenience is part of its offer, but so is the idea of local presence, of being embedded in neighbourhood life rather than hovering above it. If robot deliveries are perceived as undermining the ability of residents to move around safely, the reputational risk goes beyond a technical argument about sensors and mapping. It becomes a question about whether a co-operative business is acting in the interests of the communities it serves, or merely adopting whatever logistics solution is fashionable.

Yet it would be too simple to romanticise the pre-robot street. Britain’s pavements have long been contested, often by far larger and more dangerous intrusions than a small delivery machine. Poorly enforced parking rules, narrow footways, inadequate crossings, and a planning culture that prioritised vehicle flow have done more to restrict pedestrian freedom than any robot could achieve alone. The new machines may be best understood not as the origin of the problem but as a catalyst that forces the problem into view. If we have not decided what the pavement is for, we should not be surprised when commercial actors make their own decisions.

What, then, would a credible regulatory approach look like? Classifying delivery robots alongside micro-mobility may be a start, but it cannot be the end. The key questions are practical: where they may operate, at what speed, under what supervision, with what insurance, and with what obligations to yield or reroute. There are also design questions: how they signal their next move, how they detect and respond to mobility aids, whether they should be allowed in crowded high streets, and how complaints are recorded and acted upon. Enforcement matters too. Rules without enforcement simply invite the most aggressive operators to treat compliance as optional.

There is a temptation in technology policy to search for a neat balance: innovation on one side, safety on the other. The truth is that the balance is political. It is about whose time is valued, whose inconvenience counts, and whose rights are treated as non-negotiable. Delivery robots are not just devices; they are a proposal for how cities and towns should function, and for how the costs of convenience should be distributed.

In the coming months, the argument will likely harden into familiar camps: campaigners warning of danger and exclusion; companies promising jobs, investment and consumer benefit; ministers searching for a framework that avoids headlines about accidents while claiming Britain is open for business. But the most important voices may be the quietest: the older resident who avoids a route that once felt safe, the wheelchair user forced into the road because the pavement has become one obstacle too many, the parent with a buggy trying to pass in a narrow stretch by a bus stop.

If the robots remain rare, the controversy will pass. If they multiply, as their backers hope, Britain will need to decide whether the pavement is primarily a corridor for commerce or a sanctuary for pedestrians. That choice cannot be outsourced to an app, or deferred until after the machines are everywhere. It is already being made, one small white box at a time, on streets that were never designed for a new class of vehicle that is neither pedestrian nor car, but something in between.