Autonomous driving is not rocket science. But keeping people safe in a car is.

November 7, 2016
Holger G. Weiss
texting while driving

“Your self-driving UBER will arrive in 5 minutes” - users of the prominent mobility service in Pittsburgh have been reading this remarkable message on the app since September 14th 2016. 

This was the date that UBER launched its first fleet of self-driving cars, just weeks after a similar service was launched in Singapore - yet there still is a driver behind the wheel and it’s far away from being a mass market service. That said, the self-driving car is reality and while most people thought that it would not be a reality before 2020, traditional car makers (OEMs) and new contenders like UBER are pushing technology, regulations and consumer awareness.

Customers and experts alike are wowed by how rapidly this new step in mobility is occurring and are looking forward to safer traffic and less accidents in the years to come - yet there are still many accidents on our roads caused by problems more complex than making a car driving autonomously.

In the modern age we are all so overwhelmed by content that we are addicted to it.

Every single second we are connected to our apps, emails, messages etc. Research shows that smartphone users check their handsets up to 5,400 times per day. What’s more, many don’t stop this practice while driving a car. One could call this a “killer” feature as this behaviour is responsible for a high level of danger while driving. The numbers are shocking: the risk of an accident is 4 times higher when using your phone behind the wheel. In Germany alone were almost 400,000 car accidents with smartphones involved during 2015, many were fatal.

A major reason for these levels of distraction is the way people consuming content. Searching and Browsing on your phone is common behavior in the digital world and users continue to do this while driving. By no means is this a new problem. The European Commission launched a project named ‘gethomesafe’ a few years ago together with the German Research Center for Artificial Intelligence (DFKI). The goal was to provide a safer way to consume content and a key element of this was to use Artificial Intelligence.

Smartphones are computers and computers are multithreading and multitasking machines. They need to be trained to understand the user better in order to provide better and more personalized information. Recommendations are a good way to reduce search time. In music this has become a crucial feature.

The emergence of AI and machine-learning will  raise the quality of personalized information and recommendation .

AI also aids the understanding of how a user has previously used content and also predicts how he will use it in future. Google Now is a service that provides a nice glimpse what could be possible but still is a long way from the finished product. Machine learning and artificial intelligence will influence a lot of our decisions in the future. But how can we build AI-driven technology that makes sure drivers stay safe with their hands on the wheel or their eyes of the road? These kinds of technical challenges are not easy to solve, but with the rising tech expertise in this field there is definitively a solution.

Another major challenge we are facing regarding machine learning is data privacy. If a service provider wants to offer personalized data it can never work without a full understanding of the user.  Google Now was mentioned above as an impressive example of personalized information, because Google knows a lot about its users. Whether the data was anonymous or not, the sheer number of touchpoints Google is controlling is extraordinary. In the car context, OEMs already know a lot of their drivers, but usually they are losing control once the car is parked.

So, when it comes to cars a conflict exists between improving safety  through information and  giving up part of your privacy.

It’s the same for self-driving cars as discussed earlier. Data would be irreplaceable. It’s a main discussion point we still need to tackle particularly  in Europe.

Self-driving cars from UBER  are certainly a spectacular technology,  but all researchers, journalists, industry insiders and futurists agree that autonomous driving cars will eliminate the ‘human’ factor,  and thus the ‘human point of failure’ as well. In reality, we are a long way from all cars being autonomous. Therefore it’s fair to assume that the way content is being presented to drivers in cars would save more lives short and long term.

More featured articles