The food delivery robot was planted face down on a downtown Los Angeles footpath, its four wheels askew, when Jag Gonzalez came to its rescue.
asserted
Gonzalez → plant → rescue
Small, squat and controlled by artificial intelligence, the machine can do most tasks on its own – except stand up.
asserted
machine → control → own
“It wasn’t my first one I’ve ever flipped over,” Gonzalez said.
asserted
Gonzalez → flip → ?
“I had been seeing this abuse so much.”
asserted
I → see → abuse
Gonzalez, 28, sent the robot on its way and headed to a movie, where he told friends about the rescue.
asserted
he → send → rescue
But three hours later, there it was again on a new corner: the same droid, taking a beating from the Adidas-clad foot of a tottering pedestrian.
asserted
droid → take → pedestrian
Gonzalez tried to stop the perpetrator.
asserted
Gonzalez → try → perpetrator
But the man crouched to look the robot straight in its LED eyes.
asserted
man → crouch → eyes
“You suck,” he said – and delivered a final kick to its midsection.
asserted
he → suck → midsection
In American cities big and small, legions of similar machines are carting around takeaway orders of sushi, pizza and White Castle chicken rings and testing whether humans and robots can co-exist.
Early signs are not promising.
asserted
signs → cart → sushi
Throughout 2026, Americans have kicked, tipped and walloped the food robots – the most tangible incarnation of AI to reach their front steps.
asserted
Americans → kick → steps
Within the past year, a crowd in Los Angeles encircled and pounded a delivery robot; a man in Miami tried to hurl one off a bridge; another attacked one in the street; and in Sacramento, California, three men filmed themselves toppling and slamming a robot on the sidewalk, racking up nearly 2.5 million views on social media.
asserted
themselves → encircle → media
Some attacks are vigilantism and some may be mere knuckleheadism, the thrill of attacking the defenseless.
uncertain
some → attack → defenseless
Eunisses Hernandez, a Los Angeles City Council member whose district is swarmed by robots, said the trend had a clear message.
asserted
trend → swarm → message
“Because of how heightened people’s senses are to corporations and to the advancement of AI technology, people feel like this is the closest they get to some of that other than online,” she said.
asserted
she → feel → online
“People feel left behind.”
asserted
People → feel → ?
Roughly 5000 robots are delivering nationwide, mostly without incident, said Dan Ives, a partner and tech analyst at Yorkville Ives, an investment bank.
asserted
Ives → deliver → Ives
Over the next three years, the number is expected to swell to 8000, and analysts predict the industry will generate more than $US5 billion ($7.2 billion) in the US, Ives added.
asserted
Ives → expect → US
More than half a million robots could be operating worldwide by 2035, according to estimates from Transforma Insights, a research firm.
uncertain
robots → operate → Insights
And they could be delivering more than food as companies forge ahead with expansion plans.
uncertain
companies → deliver → plans
Serve Robotics, one of few publicly traded operators, lost more than $US100 million last year.
asserted
Robotics → trade → million
It brought in $US2.7 million.
asserted
It → bring → million
But Serve and other top companies are backed by such names as Nvidia, leader of the AI boom; Sam Altman, who helms OpenAI; Uber; and AI giant Nebius – tech heavyweights that have poured in hundreds of millions.
asserted
that → back → millions
Through partnerships with DoorDash, Uber Eats and other food delivery platforms, companies are operating in at least 10 metro areas.
asserted
companies → operate → areas
Using AI, the bots assess traffic and plan routes.
asserted
bots → use → routes
They sidle up to restaurants, pop open insulated compartments and collect orders.
asserted
They → sidle → orders
Some can stash up to six large pizzas.
asserted
Some → stash → pizzas
A variety of buzzes – texts, calls, notifications – alert customers to their arrival.
asserted
variety → alert → arrival
With bright white circles for eyes and human names emblazoned on their sides, Serve Robotics’ droids were designed to look innocuous, the company’s CEO, Ali Kashani, said.
asserted
CEO → emblazon → sides
“Vandalism and societal acceptance were very top of mind,” he added.
asserted
he → add → mind
But looks go only so far.
asserted
looks → go → ?
Some freewheeling robots have become agents of chaos, blocking ambulances, clogging walkways, colliding with trucks and trains, and along the way igniting frenzied robophobia.
asserted
robots → become → robophobia
In Los Angeles, Miami, Philadelphia and more, the robots have faced vandalism this past year, according to police departments and encounters shared on social media.
uncertain
robots → face → media
Intentionally damaging property is a crime, but arrests are few and statistics recording the reign of terror are scant.
asserted
statistics → damage → terror
At the hands of a vandal, robots are helpless, unable to call authorities.
asserted
robots → call → authorities
That makes it difficult for police to track abuse, said Jasmine Reilly, a spokesperson for the Philadelphia Police Department.
asserted
Reilly → make → Department
“Unless people report it,” she said, “we won’t know anything about it.”
asserted
we → report → it
Several companies have human supervisors who watch from afar, aiding robots in their hour of need.
asserted
who → have → need
A call to police is “very rare”, said Kashani of Serve.
asserted
Kashani → say → Serve
“We actually don’t want to be using police resources for this.”
asserted
We → want → this
…and 35 more, not listed.
The Aporia
