Machine Learning: Self-Driving Cars Will Decide Who to Save

Posted by Pauline Farris on December 12th, 2018

The advent of AI has brought numerous technological breakthroughs to our society. Modern systems such as IoT, chatbot algorithms and smartphone assistants would be a thing of sci-fi only a few decades ago. One of the more recent, trending topics to come out of AI is the notion of self-driving vehicles.

Given the idea of an AI-driven car with human passengers inside, we need to discuss the implications of this technology. Self-driving cars would coexist with human-driven vehicles, pedestrians and ever-changing traffic laws on the streets. What are the implications of this technology, and how can we discern AI morality from pre-programmed behavior?

Man vs. Machine

The famous writer Isaac Asimov coined the terms which are now norm, such as “robotics” and the three laws of robotics. In them, he discussed the morality of machines and how they are practically unable to harm their human “masters”, even at the price of their own lives. This raises an interesting topic of discussion when it comes to self-driving vehicles and their presence on the streets.

Most casual people would scoff at the idea of seeing autonomous cars on the street because of how ominous it must look to them. Just imagine seeing a self-driving car with no driver inside for the first time, going down the road with the help of its AI. It’s easy to see why most people would be apprehensive about the technology and its mass implementation, especially at this early stage of development.

However, there is also the other side of the coin to discuss, where some drivers might be unable to operate vehicles independently. Elderly drivers, handicapped individuals and others in special circumstances would be thrilled to receive the reigns of their very own self-driving vehicle – so where do we draw the moral line?

The Morality of Code

The popular HBO series “Westworld” concluded its second season a few months ago. In it, the development of AI took a drastic turn for the worse for humans and culminated with a literal robot revolution. This grim, fictional scenario represents the perfect opportunity to dissect the morality of AI and what it means in today’s developments. Most people are blinded by the notions they accepted from popular media. Robots have inherently been used as the antagonists to humanity and very rarely turn out to be “good” in the end.

Another good example is “The Matrix”, in which humanity is quite literally enslaved by machines through simulated reality that resembles our own. However, it’s worth pointing out that we are far from this level of AI development in today’s day and age. Machine learning algorithms that precedent in AI development, meaning that they start with the basic knowledge we program into them. Over time, they accept the experiences around them as new data, retrofitting them into their next operating cycle to become more efficient tools for us. There is no ominous presence of AI lurking beneath the surface.

Mistakes can happen not only in AI but in other industries as well. Design, writing, marketing and others all experience campaign failures, bankruptcy and corruption. This doesn’t mean we should turn our backs on corporate development but learn from our mistakes and move forward with more insight. Platforms such as The Word Point continue to thrive and help companies expand their international influence with localization services. Make no mistake, there will come a time where certain actions become obsolete for humans and will be passed down to AI-driven algorithms for execution. However, people will still be able to learn, earn degrees and work side-by-side with AI code to make the world a better place.

Autonomy and AI

The question on everyone’s mind when it comes to AI is whether or not it knows about morality. The answer is more complicated than a simple yes or no however. Machines are inherently designed to do the things we want them to. For example, a kitchen blender is an amazing tool which can make cooking easier and more enjoyable. It can also severely hurt someone if it’s mishandled or turned on with your hand still inside.

The point is that AI-driven vehicles do know the difference between stopping at a red light and driving on a green one. They are also capable of stopping if a human child should jump in front of the car by running across the street. However, the same rules of physics apply for AI as they do for humans – and accidents can always happen. Blaming the self-driving car for hitting someone who jumped in front of it would be paradoxical – it could easily happen to a driver as well.

However, we have learned to blame technology before we place blame on any one person. As is the case with machine learning algorithms themselves, a self-driving vehicle knows only what we program it to know. The system requires constant updates about traffic regulations, weather conditions and other driving elements through IoT connectivity. Without this knowledge, an AI vehicle is only as useful and “aware” as the passengers it transports.

Man and Machine (Conclusion)

Asking a self-driving car to choose between bumping a child and an elderly woman on a slippery road is not a feasible scenario. As we’ve mentioned before, today’s AI is far from perfect or capable of making moral decisions which would affect lives on the road. This is why self-driving technology is still in its infancy, even though breakthroughs are being made. It is also why every demonstration of the technology has been done on geo-scanned roads far away from the busy streets of New York, London or Paris.

The only way to bring a logical conclusion to the discussion of self-driving vehicles is to mass-implement the technology across the globe. If every (or majority) of vehicles become AI-driven, they will know how to handle each other on the street. However, mixing and matching human and AI vehicles on the road can and will cause mass confusion and errors on both sides. Only time will tell what the machine learning AI will be capable of in a few years. Until then, we can only speculate and keep a close eye on the many conglomerates eyeing to take their proverbial piece of the self-driving vehicle manufacture.