Authors: Anvith Potluri, Kevin Li, Kevin Nguyen, & Varad Thorat
We are living in the information age. With constant data being processed day by day, information overload, and exponential content creation, we have the sufficient amount of data, programs, processing power, and hardware to support autonomous driving. Although this is very exciting, we must take preventative measures to make sure that it is behaving properly and ethically without facing real life repercussions. As with the health industry, where software can make an impact in life or death situations, there is no difference when it comes to self-driving cars.
In order for autonomous vehicles to be in the hands of good ethics and best practices, we must ensure that it accounts for everything. When I say everything I mean everything. We should shape its model similar to defensive driving. “It’s better to be safe than sorry” is a way to put it. Instead of taking the invert risk of running that yellow light or going over the speed limit by a little, it must default into taking the safe option every single time in any given situation in order to ensure safe consistency.
We have previously stated “It’s better to be safe than sorry.” But what if we are put in a situations where “sorry” is unavoidable? How sorry should we be if presented by 2 situations we are inevitably going to be sorry for? Just like the “Trolley Problem,” it is no different when it comes to autonomous cars. With AI and humans, us humans are the architects of the AI model and ultimately could decide which algorithm to pursue in a unavoidable situation where we have to choose “the lesser of 2 evils.” Now the abstraction of ethics is at a high level and it comes back to a full circle as to what the developer of the AI system’s ethics and morals is.
It is our responsibility to do extensive testing to be confident that we have little to minimal percent error in case something DOES go wrong. And when it does, we are sure of the better choice in an inevitable situation. This of course, trickles back down to the question of “right and wrong” and moral values of the autonomous car’s source. Us. Because in the end, computers are designed to do exactly what a human designs it to do.
Sources:
- Brown (2019). “Are we Programming Killer Cars? The Ethics of Autonomous Vehicles. (Links to an external site.)” Interesting Engineering.
- Wamsley (2018). “Should Self-Driving Cars Have Ethics? (Links to an external site.)” NPR.
- Dodgson (2016). “Why Mercedes Plans to Let its Self-Driving Cars Kill Pedestrians in Dicey Situations. (Links to an external site.)” Insider.
Ethics in Computer Science 