An obsession with today's humanized robots that go viral online is unlikely to move the manufacturing industry forward.

picture alliance via Getty Images

Here’s a familiar picture for most manufacturers. An engineer stands beside a paint operator on an assembly line. Together, they carefully adjust the angle of a shoulder, test the bend of an elbow, and tweak the rotation of a wrist. With each iteration, the process refines and the outcome improves. It’s a scenario playing out right now in plants and factories all over the country. But in this case, the engineer was me, the operator was a Fuji Automatic Numerical Control (FANUC) paint robot – and the year was 1989.

An evolving story, not a new one

That experience has stayed with me – not because it was unusual, but because it wasn’t. Industrial robots first appeared in the 1950s. By the 1980s, they were already operating at scale and during the 1990s, their capabilities expanded to ever more challenging applications.

Fast forward to today and there are now more than five million industrial robots in operation worldwide – a number that continues to grow year on year. And while there’s no denying the transformative impact this is having across the manufacturing industry, we’re far from starting with a blank slate.

Despite the sensational headlines, this isn’t a new story. It’s an evolution built on decades of trial, error, and hard-earned insight. Most importantly, the lessons we learned all the way back in the 80s still ring true. In robotics, performance always outweighs appearance.

A human problem

That fundamental truth is in danger of being forgotten – or at the very least, overlooked. Right now, attention is increasingly on creating humanoid robots, with clips of machines that walk, move, and behave like people capturing the world’s imagination.

But in a manufacturing environment, all this hype can be a red herring. As Michael Cicco, President and CEO of FANUC America, told me, “Recreating the human form isn’t necessarily the most effective way to solve manufacturing challenges, particularly as advances in AI and control technologies allow robots to better sense, perceive, reason and act with intelligence purpose‑built for industrial environments. Factories are set up to do specific things, and we should be leveraging decades of experience to build for those tasks.”

In other words, there’s a growing risk of technology for technology’s sake. Of innovation measured by spectacle rather than impact. Of an obsession with humanized robots that go viral online yet are unlikely to move the manufacturing industry forward.

Dirty, dangerous, dull…dated?

The right direction of travel lies in the past. From the outset, robotics in manufacturing has been grounded in a very practical objective: take on work that is dirty, dangerous, dull, or all three. That goal hasn’t shifted. We’re still solving for safety, quality, and efficiency. It’s just now we have better tools to do it.

Today’s robots can see, sense, and adapt in ways that were once out of reach. Autonomous systems navigate factory floors, while more advanced robotics are beginning to take on increasingly complex tasks.

Even areas like general assembly once considered too variable to automate are being reshaped by machines capable of handling tasks that require strength, coordination, and precision at a scale no worker could replicate. This includes lifting and joining major vehicle components; a task that would otherwise require considerable manual effort and create new safety risks.

Form vs. function

Crucially, none of these solutions involves a machine designed like a human. In fact, if they did, they probably wouldn’t be successful in the first place. Trying to make robots look like people introduces constraints that don’t make sense in an industrial setting.

For example, a humanoid robot that walks on two legs may look impressive – but it’s inherently less stable. Industrial robots, by contrast, are designed for balance, durability, and control. They have fixed bases or move on wheels precisely because that makes them much less likely to topple over.

The same goes for arms. Robots that are configured with multiple and/or asymmetrical limbs tend to vastly outperform those limited by a more human-like matching pair. Especially in critical applications like painting/coating, welding and sub-millimeter repeatability in assembly.

“Manufacturers are navigating a significant wave of excitement right now,” Cicco says. “When it comes to robotics investments, the key questions are the same as they’ve always been: Will it improve safety? Will it increase throughput? Will it reduce defects? Will it make the job easier and more consistent?”

Built on decades, not demos

The other risk is they forget how much practical experience already exists. Industrial robotics hasn’t suddenly emerged; it’s been built, tested, and refined over decades by companies operating in real manufacturing environments not in a lab demo for TikTok.

That experience shows up in the machines themselves as well as in how they’re deployed, integrated, and continuously improved. And it gives firms a great platform from which to invent and progress.

This must also be backed by investment beyond the machines themselves. As well as expanding its US manufacturing capacity, FANUC has launched a new robotics and automation training center to help address the national manufacturing skills gap and support the shift to AI-enabled robotics. After all, the next generation of robots will only be valuable if people have the ability to use them.

Don’t chase the wrong future

Of course, none of this diminishes the importance of innovation. The momentum behind robotics today is real and it’s opening new possibilities across the industry. But when attention shifts too far towards what robots look like, it distracts from what they’re meant to do. For manufacturers, progress therefore requires laser focus on solutions that genuinely improve safety, quality, and performance – not simply aesthetics.

When I think back to that paint shop in 1989, the robot I was working alongside didn’t look like a person, and it didn’t need to. Then, like now, what mattered was what it could do and what we could achieve by using it as part of a broader manufacturing system. As Cicco points out, “On a factory floor, the conversation isn’t about whether a robot looks human. They’re asking whether it can perform.”