On a sidewalk in downtown Los Angeles, a grown man approached a food delivery robot, looked directly into its illuminated eyes, and told it, “You suck.”
Then he kicked it.
The robot, having apparently exhausted the man’s patience by attempting to deliver someone’s dinner, offered no response. There are limits, after all, to what artificial intelligence can accomplish.
According to an October 10 report in The New York Times, the encounter was hardly unusual. Across American cities, delivery robots have been kicked, overturned, and vandalized. In Sacramento, three men filmed themselves attacking one and shared the footage online with the caption “Protecting the world from A.I.”
The robots, for their part, have declined to retaliate.
For generations, we imagined a future in which machines would relieve us of life’s more tedious responsibilities. They would clean our homes, prepare our meals, and perform the thankless chores that occupied human beings before the invention of leisure. Freed from everyday drudgery, we would presumably devote our time to art, literature, and scientific discovery.
Instead, we’ve reached a point in human development where people are filming themselves assaulting rolling lunchboxes.
There is something revealing about the hostility directed toward these machines. A delivery robot doesn’t have political opinions, personal ambitions, or an especially sophisticated understanding of its surroundings. It has a destination, an insulated compartment, and a collection of sensors intended to keep it from colliding with things. Yet somewhere between the restaurant and the customer’s front door, it has become a symbol of everything people resent about the modern world.
And not entirely without reason.
The arrival of autonomous delivery robots raises legitimate questions about who controls public spaces and who benefits from technological progress. Sidewalks exist for pedestrians, including people with disabilities who already encounter enough obstacles without having to negotiate passage around an autonomous pizza carrier. Delivery workers have reasonable concerns about machines designed, at least in part, to reduce reliance on human couriers. Residents may wonder why private companies can introduce fleets of unfamiliar devices into their neighborhoods before anyone has established whether the public actually wants them there.
These aren’t trivial objections. Technological innovation has a long history of arriving with grand promises about improving society while leaving the people expected to accommodate it with considerably less influence over the outcome. The companies developing these machines may genuinely believe they’re making delivery more efficient. That doesn’t automatically establish that the sidewalks belong to them.
But none of this explains why a grown adult would look a robot in the face and tell it that it sucks.
Or perhaps it explains more than we’d like to admit.
The machine cannot revise corporate policy, negotiate a labor agreement, or reconsider the economic consequences of automation. It cannot explain why its manufacturer chose a particular neighborhood or whether anyone living there was consulted. Those decisions were made elsewhere, by people whose names most residents will never know and whose offices they will never visit.
The robot, meanwhile, is right there.
It is the most accessible representative of an enterprise whose decision-makers remain comfortably out of reach. It has become the public face of private decisions, an unusually convenient target for frustrations it neither created nor understands. Attacking one is rather like punching a parking meter because you’re unhappy with city government. The object of your anger may be conveniently located, but it has remarkably little influence over the decisions that brought you together.
Perhaps that’s part of the appeal.
The robots are small, slow, and largely defenseless. Their rounded bodies and illuminated eyes give them the vaguely endearing appearance of household appliances that have somehow secured employment. They can navigate intersections and transport meals, but many cannot right themselves after being overturned. For all their technical sophistication, a sufficiently determined pedestrian can bring an entire delivery operation to a halt with one well-placed foot.
A human courier might object, argue, or defend himself. A robot simply lies there.
There is an uncomfortable distinction between opposing something powerful and attacking something powerless, even when the powerless thing is merely an expensive collection of circuitry. The robot cannot experience humiliation, but the person attempting to humiliate it is still making a choice.
And sometimes that choice comes with an audience.
Social media has transformed countless ordinary displays of poor judgment into opportunities for public recognition. A person who knocks over a delivery robot can now present the act as political resistance, record the entire performance, and distribute it to thousands of strangers. The rebellion requires little sacrifice beyond the possibility of being identified. No organizing, no public meetings, no difficult conversations about economic policy. Just a camera and a machine that cannot run away.
There is a certain irony in using a sophisticated smartphone and an algorithm-driven platform to announce one’s opposition to technological progress. The resulting engagement is presumably of considerable interest to the same industry the protester believes he’s undermining.
But perhaps the strangest part of our relationship with these machines is how eagerly we’ve begun treating them like people, even as technology sometimes makes it easier to overlook the people performing the same work.
We give robots names, expressive eyes, and personalities carefully designed to invite affection. Meanwhile, app-based delivery services can reduce the human beings preparing and transporting our food to little more than a photograph, a first name, and an estimated arrival time. We watch their progress on a screen, become irritated when the numbers change, and occasionally forget there’s an actual person on the other end of the transaction.
The robot gets a face. The worker gets a performance metric.
There is a peculiar convenience in this arrangement. Companies can design machines to appear personable without having to grant them any of the consideration that actual people require. A robot doesn’t need a living wage, sick leave, or a sympathetic manager. It can be given a cheerful name and a pair of friendly eyes while the economic decisions surrounding its existence remain entirely impersonal.
The appearance of humanity has become another feature of the product.
For all the effort that goes into making these machines appear personable, their most revealing encounters with humanity have remarkably little to do with technology. What matters is how people respond to something that has become the visible representative of decisions made elsewhere.
Consider Jag Gonzalez, a Los Angeles resident who encountered an overturned delivery robot and stopped to put it back on its wheels. He had done this before. Later that evening, he encountered the same machine being attacked again and attempted to intervene.
The robot wasn’t suffering. It couldn’t feel abandoned, appreciate being rescued, or remember the stranger who had helped it. Gonzalez wasn’t saving a life or performing some extraordinary act of courage. He was simply responding to something that had been knocked over and could not get back up.
He saw that he could help, and he did.
What interests me about his response is that it required neither affection for artificial intelligence nor approval of the industry behind it. Gonzalez didn’t have to believe that delivery robots belonged on public sidewalks. He didn’t have to endorse automation, defend corporate profits, or celebrate the efficiency of transporting dinner without a human courier.
He merely had to recognize that the machine in front of him wasn’t responsible for any of those decisions.
That distinction is easily lost when frustration with an institution becomes hostility toward whatever happens to represent it. The robot may carry a company’s name, but it cannot answer for that company’s choices. It was sent there by people who remain considerably harder to confront.
Holding those people accountable requires public participation, regulation, and sustained attention to questions that rarely produce entertaining videos. Knocking over a robot requires considerably less effort.
We are increasingly surrounded by the visible consequences of decisions whose authors remain invisible.
The machines aren’t responsible for that distance. They’ve simply made it easier to see.
We should question the arrival of autonomous technology in shared public spaces. We should examine its consequences for workers, insist on accessibility, and demand meaningful public oversight. Technological progress should not become a justification for removing human judgment from decisions that affect human lives.
But neither should frustration with those decisions become an excuse to abandon our own judgment.
What stays with me is the contrast between two strangers encountering the same machine on the same evening. One apparently regarded it as something worth attacking. The other saw something that needed help and put it back on its wheels.
Neither changed the policies of a technology company. Neither resolved the future of automation. But each made a choice about how to respond to a world increasingly shaped by forces beyond his immediate control.
We’ve spent extraordinary resources teaching machines to navigate our streets, anticipate obstacles, and find their way through an environment they did not create.
The people making the decisions, meanwhile, have become remarkably good at staying out of the way.
Perhaps the real question isn’t whether robots will eventually become intelligent enough to understand us.
It’s whether the people responsible for putting them on our sidewalks will ever have to listen.