For all the challenges surrounding operationalizing robotics, Jan Bosch observes enormous progress in automation over the last decade, especially where AI, electronics and mechanics are combined.
One of the most interesting learnings from recent years is that it turns out that what we call “white-collar jobs” proved to be much easier to automate with AI than so-called “blue-collar jobs.” White-collar jobs require that thing that we think makes us so special as humans, intelligence, and most of the tasks involved are concerned with interacting with a computer. It turns out that many of these jobs are really not that special after all and can be automated and replaced by machine learning-based solutions quite easily.
The blue-collar jobs, on the other hand, require more than just the brains. They involve interaction with the physical world, ranging from driving a car or a truck to moving goods in a warehouse and from cleaning a toilet to constructing a new building. It turns out that these jobs are much, much harder to automate than the ones behind a desk.
The good news is that over recent years, we’ve finally been making some serious progress on robotics and our ability to automate tasks that require interaction with the real world. The limelight is of course stolen by humanoid robots, such as the Optimus presented in prototype form by Tesla, but there’s a lot more going on than meets the eye.
The first distinction we have to make is between humanoid robots that are supposed to be generic solutions replacing humans in different tasks and roles and specialized robots that are designed for specific tasks and jobs. According to one article, there are more than twenty real, high-fidelity prototypes of humanoid robots. A high-profile one is Atlas from Boston Dynamics – many of us have seen the video where it does a backflip. Other examples are concerned with robots that are used in factories or construction sites. For example, the ARMAR-6 from the University of Karlsruhe apparently can use drills, hammers and other tools. Another category is humanoid robots that are intended for interaction with humans and that add emotion to their palate of interaction, such as a bartender robot.
A related category consists of robots that are copies of animals. Many of us have seen robot dogs, but many other animals are being recreated as robots. [One article}(https://roboticsbiz.com/bioinspired-robots-top-25-robots-inspired-by-animals/) provides a nice overview of robots mimicking fish, spiders, dogs, birds, salamanders, snakes, jellyfish and even scorpions. Although often more a proof of concept than clearly intended for specific tasks or use cases, these robots are also making progress. A snake robot can, of course, maneuver more easily in tight spaces such as inspection of water lines and wastewater piping.
The specialized robots are where many fail to see the rate of progress. The primary example is autonomously driving vehicles. This problem, like most robotics challenges, proved to be much harder than what many expected. I remember talking to friends in the automotive industry about electrification and autonomous driving. Everyone expected electrification to go slow and autonomous driving to be realized quickly and it turned out, for a variety of reasons, to be the opposite. Even if it has taken years, we now have autonomous taxis without a safety driver on the road in the US, offered by Waymo, and autonomous trucks are on the road as well, although there still is a safety driver in the cabin. It won’t be much longer until autonomous vehicles are available worldwide to the great benefit of humankind. Currently, according to the WHO, more than a million people per year die in traffic-related accidents across the globe. This technology has the potential to reduce that by orders of magnitude.
Of course, the use of robots in logistics doesn’t end with autonomous trucks. Numerous companies out there provide automated solutions where pallets are automatically loaded and unloaded by robots. In combination with automatic guided vehicles and warehouse automation, this can automate the entire logistics chain for warehouses and factories.
The use of robots in medicine is growing rapidly as well. Many are aware of surgical robots, such as Da Vinci, a system that has now been used in over six million operations. There are many initiatives underway to develop dedicated robots for specific purposes such as joint replacement, endoscopy and neuroscience. Also, work is ongoing on microrobots that can be inserted in the bloodstream for conducting medical operations. A very exciting application is in prosthetics where AI, interaction with human nerves, electronics and the mechanical parts of a prosthesis are combined to create limbs that are as close to their human counterpart as possible. Incredibly exciting to follow and amazing for the people who see an enormous improvement in the quality of their lives because of these developments.
For all the challenges surrounding operationalizing robotics, enormous progress has been made over the last decade, especially where AI, electronics and mechanics are combined. Ranging from autonomous vehicles to logistics robots and from robots mimicking animals to AI-enabled prosthetics, in many areas, robots are reaching the stage where they can be industrially deployed. This allows humans to avoid dangerous, heavy and boring jobs as they can be replaced or supported by robots.
Although the Luddites among us will bemoan the loss of jobs, I feel that many of these jobs were not that exciting and valuable for humans to begin with. So far, we’ve been able to create new jobs that are better, more intellectually stimulating and often better paying than the ones that were disrupted. I see no reason why this shouldn’t be the case going forward.
Robotics and automation are yet another reason to be optimistic and excited about the future. To end with a quote from Bill Gates: “Robotics and other combinations will make the world pretty fantastic compared to today!”
