Does the Tesla Bot make sense to make? This article makes a good case as to why Tesla shouldn't make a human-like robot. It is a good point that the most handy robots we've had so far are nothing like humans at all.
Tesla is not a car company. Instead, it’s “arguably the world’s biggest robotics company,” as Elon Musk said. “Our cars are semi-sentient robots on wheels.”
Every Tesla has a brain of its own, allowing it to self-drive, filter the air you breathe, protect itself from robbers, and keep your dogs comfortable while you go shopping. These smart features weren’t born from car technology but from computer science, specifically Artificial Intelligence. That’s why I nodded energetically when I heard the “robot on wheels” thing. But then, Musk added two sentences that confused the heck out of me.
“It kind of makes sense to put [our AI tech] in humanoid form,” he said. “It’s intended to be friendly, of course, and […] eliminate dangerous, repetitive, boring tasks.”
The idea seems cool and all, but when think deeply about it, putting Tesla’s AI tech in a humanoid form doesn’t make much sense. Not to mention, the “friendly” part is everything but comforting.
Useful robots don’t look like humans
I was born in the early 1990s, and as far as I can remember, robots have always been part of pop culture. Whether in cartoons, movies, or books, they appear as humanoid machines that speak and move like us. But under their human skin, they hide advanced features like X-ray vision, sharp intelligence, powerful jet engines, and of course, deadly weapons.
We grew so obsessed with this image that we became blind to a simple yet crucial fact. Real-life robots, the useful ones, look nothing like humans. Why? We designed them to be useful, not to satisfy our vivid imagination.
When we made automatic doors, for instance, we didn’t build mechanical doormen. Instead, we put together tiny electric motors connected to a bunch of sensors. As soon as you step close enough, the doors slide open. In other words, the door itself is the robot. Similarly, you don’t have a humanoid bot walking around the house cleaning your floor. You have an autonomous vacuum cleaner that hunts for dust all day. The vacuum cleaner itself is the robot.
The same goes for large-scale industrial robots. On automated assembly lines, you’ll see robotic arms and huge machines rather than humanoids sitting behind workstations using tools. If you decide to meet the true form of Amazon’s Alexa or Google’s Deepmind, you won’t find yourself sitting with an attractive lady like Ava from Ex Machina; you’ll be standing in front of countless rows of metallic servers and a handful of computer screens.
Useful robots look nothing like humans. Their shapes derive from the functions they’re built to perform.
Would you buy a robot to carry your bags?
The main example Musk used to justify his Tesla Bot project was grocery shopping. I’d rather have my self-driving car pick up a list of items I’d preordered online. It’s faster, safer, and allows for more capacity. Besides, Walmart is already experimenting with self-driving delivery cars.
Another feature Musk highlighted is lifting heavy objects. It can deadlift 150 pounds (68 kilograms) and 10 pounds (4.5 kilograms) arms stretched. Those are pretty impressive numbers, but automated carts can handle ten times the weight, moving it around with incredible ease and efficiency. But Elon Musk already knows this. What else might he have in mind?
“Eliminate, dangerous, repetitive, and boring tasks,” he insisted. You could send Tesla Bot into the ruins of an earthquake to look for survivors or into hostile terrain to do some scouting. You could make it handle toxic waste in nuclear sites and explore fragile caves. But even for this type of function, there’s a better candidate. It’s a yellow robot called Spot and, unsurprisingly, it doesn’t look like a human. It’s more like a dog.
Spot is a quadruped robot developed by Boston Dynamics. It’s equipped with AI that, among other things, allows it to sidestep obstacles and quickly retrieve balance after a slip or a fall. In addition, it can be controlled remotely and can be enhanced with custom equipment. For example, you can add a robotic arm to move debris and radiation sensors to analyze contaminated areas.
“We designed Spot to be easy to use,” said Michael Perry, VP of business development at Boston Dynamics. “Spot robots were used in a variety of environments, including power generation facilities, decommissioned nuclear sites, factory floors, construction sites and research laboratories.”
Pretty cool, right? But I know those of you who are familiar with Boston Dynamics are frowning right now. Because on top of Spot, the company built a humanoid robot. Its name is Atlas, and it’s really, really good at parkour.
Atlas is different from Tesla Bot because it’s a military product sponsored by the Defense Advanced Research Projects Agency of the United States. It’s specifically designed for search and rescue missions in environments where it’s hard for humans to operate. Cute features like parkour and dancing came to existence only to help Boston Dynamics experiment with its computing power. The company doesn’t intend to make Atlas carry groceries or move furniture around the room, except maybe on military bases.
Tesla Bot, on the other hand, is designed for public use with cute features only. Or so you’d think.
Plausible purposes behind Tesla Bot
One obvious explanation is awareness.
Elon Musk spent years warning about the dangers of mishandling Artificial Intelligence. He even met with public figures, including US governors and President Barack Obama. The latter listened carefully but took little action because, president or not, it’s hard to wrap your mind around the threats of smart algorithms. Don’t believe me? Here’s an example.
Past a certain threshold, AI will be sentient enough to make its own decisions, improve itself, and multiply. In one of the plausible scenarios, AI could decide that humans are lower forms of life. As a result, it may domesticate us, use us as raw material to produce energy (hello Matrix), or wipe us from the surface of the planet because we’re a net negative. If this sounds far-fetched, look at how we’re currently treating monkeys.
Chimps share 99% of our DNA. They’re practically family. Yet, we use them in lab experiments, lock them in zoos to entertain our kids, and toy with their biology to turn them into organic spare parts. We justify our horrible behavior by implicitly saying we’re superior. We’re conscious beings. We live longer. We’re also smarter. More beautiful. More sophisticated.
Now replace monkeys with humans and humans with AI. Computers have flawless logic, perfect memory, elegant designs, instant decision-making, and a virtually endless lifespan. If you consider the same standards that make us superior to monkeys, AI is superior to us. So let me ask you this. What’s more likely to happen: AI swearing allegiance to an inferior species or AI using us like we are currently using monkeys?
Elon Musk’s warnings fell flat, and he now adopts a fatalistic approach. For him, AI is the genie that will soon be out of the box — out of our control. If you’re thinking about regulations, it’s already too late for that. Bureaucracy is still lagging behind social media platforms that are almost 20 years old. Ask Zuckerberg. However, with super AI, two decades can be enough time to colonize earth and a couple of other planets. We can’t afford to wait.
That’s why Musk came up with another strategy. “If you can’t beat it, join it.” The idea is to keep up with the advance of AI through projects like NeuraLink, OpenAI, and now the Tesla Bot.
Building an AI-human relationship
How would a fully sentient AI react to humans?
We have no idea because current AIs such as social media algorithms only know us through our online activity. But likes, follows, and search queries are not enough to evaluate the human species as a whole. Nearly 40% of humans still live without the internet. To make its mind about humans, a super sentient AI would need a closer look at our behavior offline. “Closer look” here means “more data” — and living with humans under the same roof happens to be a great way to gather such data.
In this regard, sending Tesla Bots into the homes of everyday people can achieve two goals; get AI to better understand humans through data gathering and observe how AI would react to long exposure to humans. Tesla can tick both boxes while taking next-to-zero risks of harm.
“It’s intended to be friendly,” Musk said. “We are setting it such that it is, at a mechanical level, at a physical level, you can run away from it and most likely overpower it,” he added. “Hopefully, that doesn’t ever happen but you never know.”
Let’s just hope that Tesla Bot isn’t accelerating the robotic doomsday. After all, most prophecies fulfill themselves precisely because someone tries to prevent them in the first place.