Emerging Technologies
Self Driving Cars
Class Topics
In this section you'll find the topics we've discussed in class and how the pertain to self-driving cars. Topics include:Â ethics concerns, privacy and network security, intellectual property, errors, failures & risks, and work and wealth
Ethics Concerns
In a brake failure, whose life do you prioritize? A pedestrian's or the passengers? This is a common moral dilemma. The car needs to be programmed with some bias as to what decision to make if human intervention is not possible. The MIT MoralMachine project highlights a wide variety of choices that a car may have to make. While taking this quiz with the group, we learned that the answers are not always so clear. We often debated for a short while what would be the best thing to do. A computer would have to make this decision in an instant, as there are numerous variations that could appear in the real world. Not all situations can be trained or programmed for.
Privacy and Network Security
The computer system on the car will need to be connected to the network in some way. This exposes it to a number of attacks. A big concern would be if an attacker were able to control the driving system through the network. In 2016 a chinese team of hackers were able to do just this with Tesla’s autopilot feature. They were able to control the car and stop it from 12 miles away. Beyond the security of the car itself, there is also the privacy and security of the rider the be concerned about. An autonomous car taxi service would need some way of monitoring their cars in case of vandalism. A natural solution to this would be to install cameras. Companies would need to ensure that these are secure. Otherwise their cars could turn into mobile surveillance
Intellectual Property
The AI that is currently being developed for self-driving cars is not able to be protected under current patenting laws. This could potentially lead to intellectual property claims between self-driving car developers. In addition to this, there are still questions about who would be at fault in the case of a self-driving car getting into an accident. In the case of a self-driving car without a human piloting it, the manufacturer would likely be at fault, while the driver may take the blame if there was one present in the vehicle.
Errors, Risks, & Failures
At the moment, American highways and roads are not optimized for a mix of self-driving and human operated cars. Driving and other conditions are not predictable and since not all scenarios can be programmed into the vehicle, accidents and other incidents will most definitely occur. Perhaps if the road contained only self-driving cars, they could communicate to one another and reduce this risk. [1]
Computer malfunctions also cause errors and risks in self driving cars. With around 30 to 100 computers in a self driving car, the room for error is quite large and although the software and technology implemented is sophisticated, there are still several areas in which these machines have their tech downfall. One for example, is the ability to operate in all weather conditions and being able to keep their sensors and cameras clean. [1]
According to a study conducted by the Georgia Institute of Technology, individuals with darker skin tones are more likely to get hit by a self driving car than those with lighter skin tones. This disparity even continues when researchers controlled for variables such as time of day or mild obstruction of pedestrians. The main factor in play is algorithmic bias - a concept in which human bias seeps into the automated decision-making systems. One famous example of such taking place is the 2015 Google case in which their image-recognition system labeled African Americans as ‘gorillas’ and in another study, it was concluded that IBM and Microsoft’s facial recognition system was more likely to misidentify the gender of those with darker skin tones, especially women, in comparison those with lighter skin tones. This is due to the information that is fed to these machines in the ‘learning’ stage of building these algorithmic systems. If these systems don’t receive data of minorities, then the result would be that they would have a harder time recognizing them when deployed. This is the reason as to why self driving cars have a disparity of five percentage points between the skin tones. [2]
----
Work and Wealth
The effects of self-driving cars could go much deeper than who is driving the car. Many experts predict that self-driving cars could deepen the socio-economic gap between the upper class and lower class. Similar to the advent of the highway system, these self-driving cars would allow people with greater resources and wealth to move further out of the cities, raising the prices to live in suburban areas, and reducing resources in the inner city. [1][3] Additionally, many white-collar jobs will still be able to do work from the comfort of a driverless car, while blue-collar jobs generally would not be able to.
​
This could also greatly reduce demands on public transport, depending on the implementation of self-driving cars in ride-sharing and taxi-like services. The reduce of demand could drive up prices, making it harder for those that rely on these services for daily transport. [2]
​
There would need to be effective regulations on the economic impact of driverless cars in order for it not to cause greater problems than just crashes.
----
1.)https://www.fastcompany.com/40490471/how-driverless-cars-could-drive-even-deeper-economic-inequality
3.)http://www.uvm.edu/landscape/learn/impact_of_interstate_system.html