Whether it's a busy employee calling for a Waymo, or a senior executive considering the profitability of adding embodied AI to the workforce, we delude ourselves into thinking that robots are "safer" than their human counterparts. Because humans are morons, right?
I'll just say it: robots are not 100% safe, and they never will be. I don't blame you if you don't believe me. Reports of industrial accidents involving stationary and mobile robots are usually anecdotal -- often because mandatory public reporting is not required, or the insights revealed are never compiled into meaningful statistical data for public consumption.
So, I can't even begin to tell exactly how many warehouse robots dropped pallets or harmed their human coworkers, how many surgical robots injured hopeful patients or how many other incidents have occurred with embodied AI. There are some numbers. For example, National Highway Traffic Safety Administration (NHTSA) reports tracked about 825 automated driving system incidents nationally by April 2026.
But even anecdotally, humanoid robots are getting plenty of bad press. In 2026 alone, academic research listed a handful of global embodied AI mishaps, including a remotely operated humanoid that frightened an elderly woman in Macau; a humanoid that embraced a dancer during a university performance in Xi'an; a robot that allegedly kicked a boy during a demonstration in Urumqi; a humanoid that caught fire after crashing into a barrier at the World Humanoid Robot Games in Beijing; and a humanoid seized by German police after supposedly touching a 16-year-old girl in Cologne. And we've all seen news footage showing Waymo vehicles turning busy city streets into parking lots, blocking traffic as if they were protesting for higher pay and better working conditions.
Robots are here now, and open, honest, detailed safety collaboration might be the key to making them safer sooner to protect the humans they were designed to help.
Steve Bigelow
Senior Technology Writer, TechTarget
I've spent a lifetime following technology, and the potential value of AI and embodied AI is indisputable. In hazardous environments, robots take over the dirty, dangerous and demeaning jobs, shielding workers from extreme heat, toxic chemicals, heavy molten metal, radiation and other industrial dangers. Using machines to carry heavy loads over long distances reduces severe traumatic injuries, such as broken bones and heavy falls. Replacing repetitive twisting, hammering or manual assembly reduces the risk of musculoskeletal disorders such as carpal tunnel syndrome, back pain and joint wear. Robots can achieve repetition and precision that humans cannot hope to match. That's why machines have been the backbone of manufacturing since the Industrial Revolution.
But the sense that robots will be "safe" is a dangerous illusion. Robots involve complex machinery that is typically extremely heavy and mechanically unstable, increasing the risk of malfunctions that could cause severe physical injury or death to humans. Sophisticated software, such as machine learning models, and network connectivity leave robots vulnerable to software errors and malicious acts such as hacking, as well as undesirable outcomes arising from the uncertainties of unpredictable or changing real-world environments. Uncertainty is risk, and the idea of robot safety is simply a euphemism for shifting risks as robots become realities in our workplaces.
We're making progress. Techniques such as fenceless guarding use sensors to protect humans, while fall mitigation forces the robot to stop or sit if a human approaches too closely. Cobots, designed for close human contact, employ force-limiting joints and elastic actuation to reduce collisions. Wireless networking uses isolated, encrypted links to safeguard data exchanges. At the same time, new standards are emerging, such as ISO 25785-1, to frame the design of dynamically stable humanoid robots.
I believe the notion of robot safety will follow the empirical trajectory of automotive safety. Every car crash is analyzed, flaws and limitations discovered, and then new materials, designs and standards evolve to mitigate those issues. People walk away virtually unharmed from car crashes today that would have needed body bags in decades past. The same thing will happen for robots.
The heartbreaking problem is cost. Every safety advancement was paid for with human pain and suffering. I shudder to think how many humans were injured and killed in car accidents to give us airbags, crumple zones and anti-lock brakes. How many industrial workers lost their lives and livelihoods before we got two-hand start buttons, emergency stops and lock-out tag-out.
The agonizing reality is that robot safety -- today and tomorrow -- will carry a similar price tag. Every laser fence, every camera, every emergency stop or kill switch, every behavioral model update, every line of reflective tape along a factory floor owes its provenance to human injury. Robots, like cars, aren't safer; rather, the consequences of their accidents are mitigated. Maybe it's a distinction without a difference.
But robot safety won't improve quickly or easily while business leaders and industries keep their robot safety issues confidential. Businesses that use robots are notoriously secretive when accidents occur. Maybe OSHA receives a workplace report when a robot arm malfunctions, or the NHTSA is notified of an autonomous vehicle accident, but there is little incentive to communicate the circumstances and share lessons learned.
I get it -- nobody wants bad press. But I wonder how quickly robot safety would advance if there were a common forum for sharing safety experiences from design and utilization perspectives. Safety Ledger is an early example of a centralized platform to systematically document injuries, deaths and property damage associated with autonomous systems and AI technology.
Today, however, such reporting mechanisms aren't universally mandated or centralized for further study and analysis. Let's hope that changes soon because robots are here now, and open, honest, detailed safety collaboration might be the key to making them safer sooner to protect the humans they were designed to help.
Stephen J. Bigelow, senior technology editor at TechTarget, has more than 30 years of technical writing experience in the PC and technology industry.
