The rise of Artificial Intelligence is a good example of how far technology has come in recent years. Incredible breakthroughs are being accomplished on a daily basis thanks to how smart our tools have become, but today’s society is increasingly receiving these advances with mixed feelings. I believe plenty of dystopias depicted in popular culture have done a great job at portraying how quickly a technology with good-hearted intentions can irreversibly change our society for the worse, and for good reason, public perception has learned to develop concern for how some of these new technologies will shape the world we live in.
This brings me to the self-driving vehicle industry. The excitement that companies such as Tesla have generated among the general public is quite unheard of. You can go online and find countless videos of people recording themselves on their first self-driving car ride experience, sharing nothing but awe and excitement. However, there is another set of individuals that have began to point out the repercussions that this technology will bring. You know what pays for the research and development of this kind of industry? Let me tell you, it isn’t private consumers buying the next generation of family cars.
Imagine a world where all vehicles are driverless. Although some years ago that might been reserved for science-fiction, it is now becoming more of a reality. For some of us, this might be a dream come true, but for a great deal of others it means the end of their careers. Self-driving cars are no-where close to being able to drive in complex city environments yet, but as we all know, the first task to be automated is the simple and repetitive one. You know what is simple and repetitive? Driving on a high-way. This kind of automation is not something of the future, it is already being deployed by companies.
According to the United States Department of Transportation, there were 1.5 million people in 2016 (in the US) in truck driving related jobs alone. This might not sound like a lot people, but I promise this will only be the first sector to feel the effects of the driverless vehicle era. US presidential candidate Andrew Yang I think paints a clear picture on the imminent dangers that this revolution will have on tomorrow’s workforce if the right measures aren’t taken. A great portion of the world’s economy relies on the transportation industry. Once there are no more truck drivers, road stops will lose most of their customers. I’m talking road-side hotels, diners, gas stations. Much of this sector is likely to experience the same work shortages. I will not go over the full extent of this domino effect, but I highly recommend listening to some of the podcasts Andrew Yang has been in. He brings up very real concerns about the foreseeable future (as well as possible solutions) regarding this issue.
Despite this, at the end of the day, technology isn’t inherently good nor bad. We can blame the companies that push this industry forward all we want, but the nature of the shift to autonomous vehicles is a rational one. Human drivers need to stop to eat, they need to stop to sleep, they get distracted. What better than to invest on a new kind of driver that works 24 hours a day? One that is never distracted from the road?
One can make the case that we should aim to keep the best interests in mind of those whose profession will soon be unemployable. After all, regulations have been put in place in the past to help people in such situations remain employed. A friend of mine once presented a really interesting counter-argument to this. Do those people actually enjoy the work they are being put through? The idea that a decent job is one where you exchange time for money has very morally gray areas. More often than not, workers in such sectors gain little to no experience over the years and are slowly forced into a position of never being able to change professions.
The argument that human drivers are better decision-making agents than machines is not one I see holding for much longer either. Machines are becoming skilled drivers FAST. They learn from millions of road-hours. More than a human could ever hope to experience in a life-time, and the learning process is not only becoming more efficient, but the gathering of essential data is too. The question is not even about efficiency anymore. In 2016, there were over 1.3 million road-related fatalities worldwide. Autonomous vehicles don’t need to be casualty free in order to justify deployment, they only need to have less accidents than human drivers do. On top of that, a machine is able to learn from hundreds of road accidents every day thanks data-gathering processes of already deployed vehicles. The question then becomes not ‘when will autonomous vehicles be safer than human drivers?’, but ‘when will humans be pronounced too unsafe to drive?’.
As drivers, we humans mostly rely on our sense of sight to be able to drive, yet our eyes only allow us for a rather narrow field of view. We attempt to make up for this through clever use of mirrors, but there are too many blind spots to ever hope to cover them all. Furthermore, our perception suffers from cognitive blind spots of its own. We can only see sharply in the center of our vision, we have a limited cognitive ability to keep track of objects in our visual field, some of us even suffer from bad eye-sight. The list goes on.
Autonomous vehicles in the other hand, have no problems with blind spots. A single 360 degree camera on the vehicle’s roof solves this issue in most cases. The ability of such a vehicle to pay attention to the road is superior to ours in almost every way. It is able to track any arbitrary number of objects in the environment, doesn’t suffer from fatigue or blind-spots in its attention mechanisms, its cameras resolution can be increased on demand. To top it off, its reaction time is considerably faster than that of humans. Given the small time frames that allow for maneuvering out of a potential crash, one could argue that passengers would benefit in terms of safety if the vehicle was to do the driving in such situations instead. This is indeed the approach Tesla has taken. Their cars often detect crashes before we humans do, and even assume control autonomously of the steering to dodge hazards for the driver. I recommend you look it up YouTube, it is fascinating.
This brings me to the ability of such machines to sense its surroundings past the limits of our own senses. Technologies such as LiDAR use lasers to reconstruct an accurate 3D model of the environment from depth information. Is it too foggy to see anything on the road ahead? Ultrasonic sensors have no problem with that. As far as the car is concerned, such weather conditions do not impair its vision. Furthermore, in the new era of the ‘Internet of Things’ we are entering, everything will be connected. Vehicles will be able to communicate with each other through radio in ways that will change traffic as we know it.
Why keep the state of things as it currently is? The future of the transportation industry is an unprecedented one, but it is not going to change the world over night. There will be time for jobs to adjust and for people to transition. On top of that, new professions will be created where the experience of today’s drivers is likely to be highly valued.
In my next entry, I will be writing about how the process to fully automated roads is likely to take place and the technologies we can expect to see applied to self-driving cars in future.