Opinion | Robots are less independent than we think

A video recently went viral showing a robot dancing to “Billie Jean,” stumbling midsong, and then being dragged off stage by a helper. Although the video elicited laughs online, there are many more videos of humanoid robots folding laundry, serving drinks, and playing competitive sports without any issues.

With all the discussion about “autonomous” robots or “intelligent” AI that supposedly think and act on their own, it’s easy to imagine a future where these technologies will replace human workers. But as someone who spent a year embedded inside a robotics laboratory as a researcher, I know that the technologies are nowhere near as independent as they’re made out to be.

When the cameras are not rolling, these machines are fragile, finicky, and constantly breaking down. Behind every polished demo video is an army of engineers fixing parts, guiding robots through tasks, and sometimes just doing the tasks themselves. In most cases, humans are remotely controlling the robots using virtual reality headsets, but you’d never know it from promotional footage that’s carefully edited to crop out those people. The robots aren’t autonomous. They’re a performance of autonomy.

The engineers and designers who keep these machines running rarely get credit. They troubleshoot, tinker, and essentially parent these systems into appearing capable. Ironically, they act as caregivers to machines that are, in some cases, marketed as caregivers to us.

This irony runs even deeper than it might seem. Researchers have long documented how certain kinds of work, including care work, get rendered "invisible." The same logic now operates for AI and robotics. The human effort behind the machine gets erased, making the technology appear effortless and autonomous.

When we believe machines are more capable than we are, human workers get devalued. Rather than mass unemployment, what we’re more likely to see is a growing class of “ghost workers” — people stuck doing repetitive, hidden labor that supports AI and robotics systems, with little recognition, protection, or pay. The threat isn’t that robots will replace us; it’s that we will get pushed into worse jobs while the robots get the credit.

But technology doesn’t just happen to us. It is shaped by the decisions, values, and choices of the people and institutions involved in building it. In that way, the trajectory of technology is not a natural force. It is a social process that is open to negotiation and redirection.

Critics may argue that pointing out robots’ limitations downplays genuine progress. But acknowledging the gap between hype and reality isn’t antitechnology; it serves the public interest. Research consistently shows that societies don’t automatically benefit from technological progress. What matters are the social and political choices that determine who that progress serves. We can support innovation while also demanding that it serve the public, not just shareholders and executives.

To that end, policymakers should regulate AI and robotics marketing claims the same way we regulate other industries. They should require transparency about the human labor involved in operating and maintaining automated systems. And they should fund research into the social and labor impacts of automation, not just its technical capabilities.

The public should be skeptical of demo videos and inevitability rhetoric. They can ask who benefits from the narrative that machines are taking over. And they should recognize that the people doing invisible labor, in all industries, deserve visibility, protections, and fair compensation.

Journalists and others in the media meanwhile can support those efforts by covering the workforce behind the robots and AI technology, not just the technology. The real story isn’t in the demo — it’s in what’s happening behind the curtain.

What shapes our future won’t be the machines. It will be shaped by us and how we build, deploy, and talk about the machines in our midst.