When I was a social worker, I had a young client who had learned his way around cars. Whatever you needed to do on your old pickup truck, he knew how to redneck it, as he put it. It was the first time I’d heard the term redneck used as a verb, but it turns out that’s a real thing. Seemed to me he had a future as a mechanic and that’s what I encouraged him to do. According to the client, however, there was a big problem: He in no way, shape, or form knew his way around computers.
Gearheads have become computer nerds because today’s cars have a sophisticated array of advanced computer technology in them, starting with the engine control unit (ECU). The ECU regulates a number of things, including air-to-fuel ratio, ignition timing, and boost. With a computer, you can access the ECU and flash it to improve your car’s performance.
The ECU is a basic precursor of what’s ahead. Once cars are automated, they’ll be able to communicate with each other, and they’ll be able to communicate with other bots. Google, NVIDIA, comma.ai, and NAUTO are among the companies incorporating AI with self-driving cars.
Comma.ai’s AI starts with a human driver and learns through LIDAR data and a convolutional neural network (CNN). NAUTO does away with the expensive LIDAR and feeds a neural network data from dashcams, motion sensors, and GPS. NVIDIA’s AI employs sensor fusion and deep learning through a deep neural network. NVIDIA’s Drive PX 2 supercomputer controls the software for the autonomous vehicles participating in Roborace.
Roborace, just like it sounds, is a series of races between self-driving AI cars. All machine parts being equal, Roborace is about determining which AI is the best. The AI that controls the car is the driver ”not figuratively, literally; the National Highway Transportation Safety Administration (NHTSA) has ruled that AI is legally the driver. At least, that’s what the NHTSA said in a letter to Google.
The people who drive cars are the ones who learn how to work on them. Since AI is driving them now, we can expect AI to work on cars too. We’re not very far off from seeing AI mechanics working on AI cars. Why wouldn’t these AI mechanics work on all other cars as well? Compared to the complex matrices involved in driving on the open road, operating on a vehicle will be child’s play for AI because there are less variables.
The gearhead of the past has become the computer nerd of the present, but in the future, this gearhead will have to program gearhead bots. Knowing how to program deep neural networks, convolutional neural networks, and deep belief networks is a far cry from knowing how to overhaul a transmission or grind down some brake pads.
Right now, the jobs most at risk due to automation are telemarketer, loan officer, and cashier. Taxi driver is not far behind, with an 89 percent risk of being automated within the next 20 years. While the Future of Employment identifies first-line supervisors of mechanics as being at low risk of automation, PricewaterhouseCoopers found that more male-oriented jobs are at risk than female jobs. Automotive technician, for the most part, has traditionally been a male job. Goodbye suspicious guy who says something is wrong with your car and it’ll cost you $5,000. Hello robot that ”if it’s going to be dishonest ”would have to have inaccuracy incorporated as part of its algorithm.
Let’s return to the NHTSA’s letter to Google about AI legally being the driver of a self-driving car. Although it’s a seemingly easy way around the liability issues posed by self-driving cars, it presents the dubious concept of the car as a legal entity. If the car is an entity, then by logical extension other AI bots are too. That’s nowhere near official, but it’s worth considering, especially as we gear up for all these robots to take our jobs.
My former client who wanted to be a mechanic is a millennial in danger of having his future profession outsourced to a robot, a robot that simply has to be manufactured in a manner of months to potentially be classified as a legal entity (with all the same rights as a human being?). One thing this client expressed to me is a desire to own a car. Yet, when it comes to millennials and debt, he’s part of a trend: either millennials aren’t buying cars ”they were 30 percent of the population in 2010, but they only purchased 17 percent of the vehicles. Or, they’re buying cars, and relying on financing ”2016 saw them make 28 percent of car purchases, but 63 percent of them are deep in debt.
The question is how will millennials saddled with debt be able to maintain payments in the long term if their jobs get usurped by robots? Author of The Future of the Professions and Tomorrow’s Lawyers, Richard Susskind says, I don’t think anyone can do long-term career planning with any confidence. Basically that means no jobs are safe from robots. Even creatives aren’t on secure footing.
Many millennials may need to go back to school for AI-centered jobs that teach them how to work with big data. Harvard Business Review points out that we have to start preparing the next generation for AI jobs now. Kids need to learn computer science skills early, skills that help them learn how to develop waveform analysis protocols, machine learning protocols, and encryption programs.
But how are we going to teach today’s mechanic to be tomorrow’s robotics engineer, tomorrow’s data scientist, tomorrow’s deep learning software engineer? Moreover, how are we going to convince millennials, who are already burdened with debt, to go back to school and accrue more debt in pursuit of an AI-savvy future?
The rise of the robot workers will put human ingenuity, sympathy, and adaptability to the test. Thankfully, so far we’ve proven we have plenty of all these qualities in spades. Let’s hope they stand up to the robots.
