Players, Challenges, and Implications of the Self-Driving Car Industry
One of the oldest and wealthiest industries is developing artificial intelligence to take your car’s steering wheel away from you. The way that the automotive industry handles how machines interact with people and their surroundings will have an important impact on our relationship with artificially intelligent systems moving forward. In this article, I will provide a brief characterization of the autonomous vehicle industry, discuss some technical challenges that it faces, and then turn to the implications that the technology has for us.
Since January, I have worked with BRAIQ as a business analyst intern, where I have learned what the landscape of the autonomous vehicle industry and the technical challenges that the industry faces look like. BRAIQ is an automated driving startup that personalizes the ride of self-driving cars. They do it by collecting biometric information about passenger comfort during the ride with pre-existing sensors, sending this data to the cloud where it is aggregated and analyzed, and then personalizing the automated driving system according to their preferences. It is simple from a high level, but it is actually a highly specialized hard tech challenge. BRAIQ is a team of PHDs in biomedical engineering, bioengineering, and psychology that is uniquely suited to take on this kind of project.
The Self-Driving Car Industry
First, let’s talk about the industry. Traditional OEMs have spent billions on acquiring companies to form the core of their automated driving teams. For example, GM acquired Cruise Automation for more than $1bn and Ford invested $1bn in a joint venture with Argo AI. These companies are trying to acquire and file patents in order to lock their place in as a player in the autonomous vehicle field for years to come and keep new players out. They want to make their autonomous vehicle technology the standard so that they can influence the direction that the field heads in.
Regarding emerging electric vehicle manufacturers, Tesla is the only one that I see as a real player right now. They have a significant number of patents, and the market has recognized their potential. They are now valued similarly to the biggest OEMs. However, this seems to be reflective of Wall Street’s overvaluation of tech firms considering that Tesla’s costs continue to grow and they sell millions less vehicles per year than practically every major OEM.
Non-traditional automotive companies like Uber and Alphabet-owned Waymo are positioning themselves to provide the automated driving software for OEMs to use in their cars. It is in this category of companies that we have seen the first IP ownership battle emerge, and I believe that this is just the beginning of companies going to war with each other to secure their position in the autonomous vehicle landscape.
Then, there are a number of startups that are developing automated driving technology, such as Drive.ai, nuTonomy, and BRAIQ. These companies will either partner with OEMs to have their technology be used in their cars or be acquired by them.
Now, let’s turn to talk about some technical challenges that the autonomous vehicle industry faces. The following capabilities are going to be necessary for autonomous vehicles to handle a seemingly infinite number of edge cases that SAE Level 5 vehicles will face.
- Emotion recognition. In order to feel comfortable in a self-driving car, it is going to need to be able to know what driving style you like. Right now, one of the biggest complaints regarding autonomous vehicles is that their ride is unnatural — it feels like you are being driven by a robot. Moving forward, autonomous vehicles will use sensors to understand your response to the ride and be able to adjust accordingly.
- Communication with humans. When there is a bottleneck in the road, or when cars are stopped at an intersection and there is not a clear right of way, people usually use hand gestures to figure out how to handle the situation. So, unless there is ubiquitous V2V technology, autonomous vehicles will need to be able to understand what people other than the driver want them to do.
- Decision-making based on ethics. When a pedestrian crosses the road unexpectedly, how does the car react? Really, it comes back to the classic trolley problem. I expect a utilitarian decision-making system to be implemented. There are, however, important questions that must be addressed including how we are defining utility, and whether cars should make decisions based off of the evaluation of specific actions or general rules.
- Evaluation of objects. Debris like nails or broken glass can pop your tire, but a plastic bag or water bottle will not. In order to provide a smooth ride, AVs will need to be able to accurately distinguish between threatening and nonthreatening objects.
The way that automated driving companies handle these questions will likely be the protocol for how AI handles them more generally moving forward.
If an autonomous vehicle mishandles one of these edge cases or is involved in a fatal accident, people will lose faith in them since autonomous vehicles are held to a higher standard than human drivers. Driving is a dangerous activity as it is, and people are afraid of putting their lives in the hands of artificial intelligence, even if autonomous vehicles are better at driving than they are.
A good case study for how this higher standard can be handled is the review process for airplanes by the National Transportation Safety Board (NTSB). Flying in airplanes can be scary for obvious reasons, and because of this the industry responded by having airplanes face a highly scrutinous review process, even though fatalities in airplanes are far rarer than fatalities in cars. After malfunctions and crashes, the NTSB carefully analyzes what went wrong and responds accordingly. Having a similarly extensive review process for autonomous vehicle accidents would increase trust in the new mode of transportation.
Further, the rollout of autonomous vehicles will be a gradual transition that takes place over the next few decades. The transition has begun with features such as adaptive cruise control, lane keep assist, and automatic parking. Next, augmented driving features will be introduced into autonomous vehicles which will allow them to change lanes and let drivers know when they are about to mess up. Fully autonomous vehicles will only be introduced once people are sufficiently comfortable with augmented driving, and people will still probably be able to take over when they need to. After their initial introduction, it should take about a decade for vehicles at SAE Level 5 to become widespread.
I’ll also note that the adoption of autonomous vehicles will probably be influenced by insurance companies offering lower rates to people who own them.
The automotive industry is one of the first to have billions to spend on the development of artificial intelligence. In order to stay competitive and offer customers the driving experience that they will soon demand, car manufacturers must develop this technology or be willing to use other companies’ automated driving tech in their cars. The arrival of autonomous vehicles will be the most direct and visible impact of artificial intelligence that we have seen so far. I see this as important to our understanding of our place in the world for two reasons. First, we will see that many activities that we take part in are replaceable by machines. Driving is just the beginning of our lives becoming increasingly automated and it will signal a new technological revolution. Second, it will be the first major increase in free time that is caused by artificial intelligence for most people. The question that arises is, what are people going to do when these kinds of mundane tasks have been replaced by machines? People will either pursue their passions and interests with this free time, or they will spend it using advanced entertainment technology. When traditional wage-paying occupations and routine activities are done by machines, people will have the opportunity and challenge to actualize their true self. Driving your car is a good distraction from thinking about this kind of stuff, but it will soon be taken away.
This article was written by the author in his personal capacity. The views, opinions and positions expressed by the author are his alone, and do not necessarily reflect the views, opinions or positions of BRAIQ or any employee thereof.