Frontpage‎ > ‎Blog‎ > ‎

Are we ready for self-driving cars?

posted May 4, 2019, 9:45 AM by KyungHyun Cho   [ updated May 4, 2019, 10:54 AM ]
Last Monday (April 29), I had an awesome experience of having been invited and participating in the debate event organized by the Review and Debates at NYU (http://www.thereviewatnyu.com/). By being born and raised in South Korea, I can confidently tell you that i cannot remember a single moment where I participated in any kind of formal debate nor a single chance in which i was taught how to make an argument for or against any specific topic. My mom often tells me I draw way too gloomy picture of Korean K-12 education I had, but it is true that our education system (at least when I was in it) never encouraged students to express their opinions and never taught us how to do so properly. 

I digressed. Back to my point, it was my first time participating in such an event and I was quite nervous. To further this, I was travelling for four days before this debate and had absolutely no time to prepare except for an hour in this airport, another hour in that airpot, another hour on this plane and yet another hour on that plane. Also, I was asked to oppose the statement "we are not ready for self-driving cars," which frankly sounded a lot like biting a poisoned apple. The opponent in this debate was a dear colleague, Prof. Vasant Dhar, which did not help with my anxiety, because he is a great speaker, debater and thought leader in the field of data science and automated decision making.  

I was supported by an amazing student from NYU Abu Dhabi, Muhammed Ali, who I only met right before the debate. Muhammed is only a sophomore at NYU AD, majoring in Philosophy and Economics. I was impressed by his quick grasp of the points made by participants, which included me, Vasant and Vasant's student supporter, Ankita Sethi, and logical and enthusiastic response to or support of them. 

It was a great joy to participate in this debate. Although we, Muhammed and I, lost the debate eventually to Vasant and Ankita, decided based on how many in the audience changed their mind over the debate, I learned a lot from this debate, listening to Vasant's and Ankita's arguments, and learned how awesome these students here are in thinking about and discussing these societal issues. Lucky y'all studying here. 

Anyhow, I at least prepared an opening argument before the debate and would like to share with you here. I've also attached at the end another, follow-up piece that I prepared but could not find an appropriate moment to read out in the debate. Apologies in advance as both pieces are pretty rough, because they were just a scrap material i've prepared for making a speech.

What do you think? Are we ready for self-driving cars? 

(Of course, don't forget this is an opening argument of a team who argued we are ready for self-driving cars and who lost the debate. )

===

There is a misconception that a new technology is introduced to a static, frozen society. Under this misconception, no new technology is ever ready. A reality is that the society and new technology interact with each other and evolve together to be more compatible with each other. There is no perfect time to introduce a new technology, especially when it is disruptive, revolutionary and potentially has a long-term effect on the society. In this sense, we are either ready from the very beginning or never ready, and I do not believe the central question to this debate is debatable to start with.

Rather, I believe we must carefully dissect the statement “we are ready for self-driving cars” along multiple axes. 

First, we must consider different subsets of the society. Within this axis, we must further consider different ways in which the society could be partitioned. 

One particular axis I want to emphasize is the primary purpose and use case of motor vehicles. Some use cars to drive around on their own or for their family. Some use cars to drive passengers picked up from the street, such as taxi drivers. Some use cars to deliver various products from one place to another. Some use trucks to deliver products from one location to a remote location, often taking days to arrive at the destination. 

Across these use cases, what we notice is there is a large variety of drivers according to their skills, their dedication, # of hours spent driving and other probably seemingly unnoticeable properties. Furthermore, there is a large degree of differences across different types of driving. For instance, we won’t be able to easily compare the experience of driving a taxi in New York City against that of driving a long-haul truck across North America. Driving your kids to a school in a suburban area cannot be compared to driving from Downtown Manhattan to JFK in the rush hour. 

Of course, these are only a small number of use cases of motor vehicles that i have listed. With these fine-grained categories of “driving”, which use case do we have in our mind when we ask ourselves whether we are ready for self-driving cars?

Let us consider a hypothetical situation in which we let a machine drive on its own only on highways. For instance, a human driver will drive a long-distance truck to the city boundary, a machine would take over control, drive the truck all the way across the continent and give the control back to potentially another human drive in the destination city. Certainly in this case, the difficulty of self-driving is significantly lower than that required for a machine to drive in a crowded city. Would we declare ourselves ready for this situation? If so, we are perhaps already ready for self-driving cars.

The main topic of this debate is thus ill-defined, and the validity of this statement is subjective at best. In this regard, i argue that we are ready for self-driving cars, because there are certainly scenarios under which self-driving could be deployed and used.

Second, we must consider different levels and axes of self-driving technology. Already the Department of Transportation and similar governmental authorities all around the world have begun investigating and defining what we mean by self-driving technology. What they along with many other policy makers, policy advisors and scientists have realized is that it is not a single chunk of machinery, that is self-driving technology. It is rather a large set of relevant technologies, such as sensor technology, motor technology, control algorithms, software integration and so many others, that together constitute self-driving technology. In other words, we must consider different ways in which a subset of these technologies are combined to form one instance of self-driving technology. 

At this point, we must ask ourselves a few questions in order to clarify ourselves of what we mean by self-driving technology before asking ourselves whether we are ready for such technology.

Is automatic transmission considered autonomous driving technology relative to manual transmission? 

Is self parking considered autonomous driving technology? If so, is the society ready for this feature? Ford, BMW, Volkswagen, … all have actual products in the market that support this technology.
 


At what point have we begun considering ourselves to be ready for some of those technologies above? Similarly to the earlier argument i made, it is an ill-posed problem to ask whether we are ready for “self-driving cars”. If there is any hope in answering this question, which needs to be better specified, perhaps the criterion should be that whether such a technology, or its variants, has already been deployed and used successfully in the society. If we agree such a criterion is a reasonable one, i argue that we are and have been ready for self-driving cars for long time.

In fact, what I argue here is that we have already been ready for these new technologies, as this readiness to adopt new technologies and even new ways of thinking is precisely what differentiates ourselves from other species. 

=== 

At this point, I hope I have been able to convince at least some of you and to establish that it is impossible to argue for or against our readiness in adopting self-driving cars. We have been ready for these kinds of technology for many decades, and gradually have been implementing these technologies in the real world. Self-driving “evolution” not revolution has been under underway ever since the successful demonstration of self-driving van by Pomerleau in Pittsburgh in late 80’s and early 90's, and has become much more evident exactly a week ago when Tesla demonstrated their latest self-driving capability.

I do not want however to give you the impression that everything has been solved and that our roads will be full of self-driving cars soon as in a couple of years. I also am not arguing that we should just see what happens.

Each and everyone of us is ready for self-driving cars on the road, but we do not necessarily agree on the timeline of their deployment, how we would and should regulate them just like how we have done so with various other technologies (including human-driven cars!), and their future. These decisions cannot be made before self-driving cars hit the road (though, they have already done so) and will have to be newly created and constantly amended over time as we deploy more and more of them across the society.

An interesting example of a similar spirit happened decades ago in Sweden, when they overnight switched from driving on the left (just like in London) to driving on the right (just like in New York City.) this decision was made many years after cars had become ubiquitous on the streets of Stockholm, and was made based on the government’s conclusions drawn from observing how cars are used and driven on the roads. What this implies is that it was not whether Swedes were ready for the technology but what Swedes could agree to be a better way of using this new technology of “human-driven” automobiles. 

We call autonomous driving a new technology despite its age. We talk about our readiness and technology’s readiness because it is a new technology. We are worried and excited about this new technology. At the very root of these different “feelings” about the new technology of autonomous driving lies the lack of our informed imagination of how the new technology would change our society and how our society would change the technology. 

Can we then wait until we have a clearer picture of the future? Unfortunately i do not believe it is possible to do so. Our ability of looking far into the future is simply not there. In other words, we are incapable of making the optimal action, and we must acknowledge that. We however have an ability to look into near future and adapt ourselves and our society accordingly.  And, there are a few problems we can make an educated guess and hopefully are working toward fixing them, such as value alignment problem, liability issues leading to potentially deepening societal inequality and economic incentives (from manufacturers and operators) leading to economic issues. All of these have been correctly and appropriately raised by Vasant and Ankita earlier today. 

Then the question is whether we should keep self-driving cars away from roads because we would never be ready for them (though, again, self-driving cars are already on the road) or we should embrace the reality that self-driving cars will be deployed increasingly more over time and that we are ready to adapt ourselves and our society to be in harmony with these self-driving cars. I believe we are ready. The next and perhaps better question to ask ourselves and debate about among ourselves is: “are we preparing ourselves to adapt ourselves and our society to the ever increasing adoption of autonomous technology for our benefit?” 




Comments