Artificial Intelligence in everyday driving school life

Autonomous driving without artificial intelligence? Not possible! Prof. Dr. Stefan Gröner talks about how far the technology has come today and what might be possible tomorrow.

Driving School: What exactly is meant by artificial intelligence? Prof. Dr. Stefan Gröner: The research field of "Artificial Intelligence" (AI) tries to replicate human-like intelligence. Through programming, clear instructions are given to solve problems defined by humans. And how will AI change our daily lives in the future? I assume that our daily lives will soon be extremely interconnected - think: Internet of Things. For example, if you search online for a recipe for eight people using a Thermomix, the refrigerator will automatically send a shopping list of the ingredients missing for that dish to your smartphone. At the same time, your calendar will be synchronized with the next major event with guests, and an offer for online purchase and delivery of groceries will be prepared in time for the event. In general, technological developments that make life easier for people will eventually prevail. In which areas is artificial intelligence still being used? In many areas. In our personal environment, examples of artificial intelligence are Apple's "Siri" or Amazon's "Alexa". Through algorithms and high processing power, speech input can be contextualized, and the most suitable answer can be given. The basis for this is having a large amount of data. Companies that generate as much data as possible in the future and intelligently analyze it will have the best offerings. The more data available, the more accurately the programs can provide relevant results through machine learning. Machine learning? These are programs capable of recognizing patterns and regularities, learning from them, and drawing new conclusions. Knowledge is generated through experiences made by neural networks or computers. For example, facial recognition programs can already learn and recognize certain personality traits. However, the technology still needs to be refined, as error susceptibility and the possibility of manipulation are still very high. As of today, how intelligent is artificial intelligence? Currently, it is as intelligent as the corresponding program written by humans. We are far from machines becoming more intelligent than humans and evolving without human intervention - as some fear. AI programs precisely reflect what is predefined by humans. For example, there is a board game called Go. The AI program AlphaGo can now defeat very good players. However, if the rules or the game board are slightly altered, humans can easily adapt, but the program no longer knows what to do. Where does artificial intelligence still face limitations? Usually when unexpected situations arise. The big challenge is to enable AI to handle such situations. Especially in the field of autonomous driving, AI plays a significant role. The technology works very well on roads with few disturbances, such as highways. Google's autonomous cars have driven several million kilometers accident-free, but only on roads with largely standardized conditions. The situation is different in urban traffic: children suddenly run onto the street, a paper bag flies in front of the car, or there is a construction site on the route - these are still challenges for the systems. There have been experiments where traffic signs were slightly altered by painting them with color. Any person can recognize the traffic sign despite the change, but the software in the self-driving car may not. As a result, the car stops. How long will it take until such situations are no longer a problem? It is difficult to predict. However, I believe that in a few years, we will already have autonomous driving in certain areas - such as on highways. It will take longer in dense city traffic during rush hour. But eventually, the technology will overcome this too. The more data and patterns available, the sooner systems with the appropriate computing power can make the right decision in various situations. In what areas does AI influence the professional world of driving instructors? Especially in the field of autonomous driving. The computer will take over tasks like reverse parking or highway driving in the future. As a result, normal traffic will play a decreasing role in training. Instead, challenging driving and safety aspects will become more important, as well as flexible and individual mobility. Driving schools should consider diversifying into mobility education. A solid education in terms of mobility will remain important in the future, but it may not be limited to just cars anymore. Such changes also offer opportunities. How will our mobility behavior change - also due to autonomous driving? Entertainment and connectivity in the car will become more important. The more driving functions the systems take over, the more crucial aspects like communication and the use of various media offerings become. This means manufacturers must provide interfaces for video and music services. In my opinion, the overall package of a car, its connectivity, media offerings, and autonomous driving capabilities will become increasingly important. Aspects like engine power will lose significance. Automobile manufacturers must rethink their strategies if they want to succeed in the market. In 10 to 20 years, engineering alone will not be the main reason why people choose a particular model. Furthermore, there will be new forms of mobility. What are those? I am convinced that in the future, we will move in three dimensions. In addition to autonomous driving, autonomous flying will become prevalent. Means of transportation will be small drones that can accommodate two to four people. In Dubai, there are already plans for such bike taxis starting in 2018, which will transport passengers from point A to point B. Mobility will significantly evolve, especially in urban centers.