Can cars tell what
people are thinking?

The Archive of the Future asks intriguing questions about innovation and technology – and selected experts respond. Beginning in September 2021, we began publishing exciting answers to fascinating questions in five issues of our newsletter. They are all collected here for your reading enjoyment.

Prof. Dr. Alexander Waibel is a leading scientist in the field of artificial intelligence and a member of the National Academy of Sciences Leopoldina. One of his AI projects has been a simultaneous translation software program that was acquired by the video conferencing provider "Zoom" in 2021.

Prof. Waibel, can robots dream?

* "Did you know that the phrase 'I have to sleep on it' exists in almost every culture? There's a reason for this: we know that we process and solve problems while dreaming – even though it's not entirely clear to us how this works. In the so-called Rapid Eye Movement phase of sleep, any external, physiological coupling of the brain is switched off. While nothing can penetrate from the outside, the emotional centre of the brain races along unhindered. Dreaming is a patently internal process. Not unlike within the human brain, AI consists of neural networks that can be trained. So one might naturally assume that artificial intelligence is also capable of dreaming. But so far this isn’t the case because dreaming relies on the presence of emotions, which AI does not yet possess. In addition, the so-called episodic memory that enables the remembrance of one's own experiences is absent. AI cannot recall today the things it did last evening. But there is one level within the brain that is more than capable of creating something new entirely on its own: I'm speaking here of the so-called Hidden Layers. These layers develop in AI because although AI can be trained, it isn't taught what to learn internally. Projects such as Google's DeepDream have used data from these hidden layers to create images that have never been learned in this specific constellation before. These images are reminiscent of human dreaming, especially with their psychedelic visuals. So AI can certainly generate exciting material. But is this artificial creativity a counterpart to humanity? I don't think so. Having said that however, it's important to note that creative AI’s potential is far from having been exhausted at this point."

Dr. Peter Spork is a science writer and biologist. In his book "The Measurement of Life," he focuses on the potential of systems biology and how it could revolutionize medicine in the future. He explains how our "digital twin" can predict the future and which questions it might well answer for us.

* "Imagine having a navigation system that monitors your health: it could calculate the conditions under which you would reach your destination and when. Life is much more complex than getting from A to B of course but the basic framework of the digital twin is the same. It is grounded in formulas and algorithms from systems biology. This involves analyzing living processes to predict how life might evolve in the future. So I can then ask my digital twin: What might my life look like in ten years' time if I start jogging every evening starting today? Or if I decide to drink two beers every night from now on? The twin then simulates the future - and thus increases my personal freedom in being able to define conditions in such a way that allows me to lead the life I want to live. Combined with data science, systems biology will take on many tasks currently burdening women physicians. This will allow medicine to move toward prevention and away from just treating symptoms. However there is still a long, continuous process ahead of us before we all have a highly complex, digital avatar full of formulas and algorithms at our disposal. Experts think it will take at least 30 years for this to develop. Nevertheless, we need to establish the course now to ensure that this vision remains a positive one – because the potential for abuse when dealing with sensitive health data is high. So if we start handling our data responsibly now, the digital twin could become a veritable utopia."

Dr. Wolf Singer is a brain researcher and one of the world's leading neuroscientists. He researches the complex networks that make up our brain and can explain how human creativity is generated.

Dr. Singer, how are inventions discovered?

* "The human brain is a purpose-driven organ. Its main attribute lies in being able to create models of the world that equip people for the future – because those who have already thought ahead are better able to react in a crisis. This is made possible due to inter-neuron communication, which also generates ideas and linguistically formulated thoughts. So what are the conditions the brain relies upon to be creative? Creativity occurs when new connections are made between neurons. This is able to happen as a result of highly complex activity patterns that are always on the move within this network of neurons. At some point, they form a link between two points that were not previously connected. This is how something new can come into being. The creative act, this ‘thinking about it,’ can be spurred on by focusing very intensively on a subject while also taking breaks to allow the brain to run free, which it continues to do at a subconscious level. By taking adequate breaks, the brain is able to engage in internally-organized processes. The basis for this emerges during early childhood development: The human brain continues developing until about the age of 20. The earlier your mind initiates these imprinting processes, the more these connections can be relied upon to develop and mature. It's important to give the brain enough to think about early on so that it has time to mature and grow."

Dr.-Ing. Bernhard Müller is the Fraunhofer-Society's spokesperson for the Additive Manufacturing Field of Competence. Based in Dresden, 3D printing expertise is gathered and integrated from across 19 partner institutes of applied research under the umbrella one of the world's leading research consortiums. We asked him if 3D printing technology can now fulfill our every whim, any time, any place?

* "Freedom of design is almost unlimited within the field of 3D printing. Manufacturing can now be done in an entirely different manner: Rather than in factories in the Far East, for instance, where large quantities of goods are produced and then shipped across the globe via container, manufacturing can now take place where products are needed. This technology offers enormous potential for the improved distribution of complex products. In many fields of medical technology this has already become the new norm. It’s still not enough though to just plop a printer down in a far-flung place and press ‘print’ in order to create a highly complex commodity such as an implant. 3D printing demands craftsmanship, engineering and expertise. We also have to bear in mind that a 3D printer is not some kind of ‘black box’ that magically ‘conjures’ the wares we need. In reality, there are many highly complex physical processes at work within a printer, which call for ongoing development and know-how. Things tend to work a little differently than on the Starship Enterprise. But will these types of processes one day improve people’s living conditions? Yes, I think so, yet we've still got a number of hurdles to contend with before that day arrives. Research is driving several issues right now, such as maximizing resource efficiency with respect to 3D printing or customized product adaptation. Currently we are also innovating in the area smart materials that will enable products to change shape in reaction to external influences. 3D printing is a decisive technology when it comes to programmable materials. A lot is going to happen in this area in the next few years."

How does autonomous driving work? Will we all one day have our very own personal KITT, the talking car from the "Knight Rider" series? Dr. Patrick Gebhard is head of the Affective Computing group at the German Research Centre for Artificial Intelligence (DFKI). He researches methods for making communication with technical systems more sociable.

Dr. Gebhard, can cars tell what people are thinking?

* "Imagine having a navigation system to monitor your health: it could calculate the conditions under which you would reach your destination and when. Life is much more complex than getting from A to B, of course, but the basic framework of the digital twin is the same. It is grounded in formulas and algorithms from systems biology. This involves analyzing living processes to predict how life might evolve in the future. So I can then ask my digital twin: What will my life look like in ten years' time if I start jogging every evening starting today? Or if I decide to drink two beers every night from now on? The twin then simulates the future - and thus increases my personal freedom in being able to define conditions in such a way that I can lead the life I want to live. Combined with data science, systems biology will assume many tasks currently burdening women physicians. This will allow medicine to move toward prevention and away from just treating symptoms. However, there is still a long, continuous process ahead of us before we all have a highly complex, digital avatar full of formulas and algorithms at our disposal. Experts think it will take at least 30 years to develop. Nevertheless, we need to establish the course now to ensure that this vision remains a positive one – because the potential for abuse when dealing with sensitive health data is high. So if we start handling our data responsibly now, the digital twin could become a veritable utopia."

Dr. Wolf Singer is a brain researcher and one of the world's leading neuroscientists. He researches the complex networks that make up our brain and can explain how creativity is generated in the brain.

Dr. Singer, how do you find inventions?

"The human brain is a purpose-driven organ. Its most beneficial function lies in creating models of the world that equip people for the future - because those who have already thought ahead are better able to react in a crisis. This is the result of communication between neurons, which also generates ideas and linguistically formulated thoughts. So what are the conditions the brain requires to be creative? Creativity occurs when new connections are made between neurons. This is able to happen as a result of highly complex activity patterns that are always on the move within this network of neurons. At some point, they form a link between two points that were previously unconnected. This is how something new comes into being. The creative act, this ‘thinking about it,’ can be spurred on by working very intensively on a topic while also taking breaks to allow the brain to run free, which it continues to do at a subconscious level. So by taking adequate breaks, the brain can engage in internally-organized processes. The basis for this is created during early childhood development. The human brain continues developing until about the age of 20. The earlier your mind initiates these imprinting processes, the more these connections can be relied upon to develop and mature. It's important to give the brain enough to think about early on so that it has time to mature and grow."

Dr.-Ing. Bernhard Müller is the Fraunhofer-Society's spokesperson for the Additive Manufacturing field of competence. Based in Dresden, 3D printing expertise is gathered and integrated from across 19 partner institutes for applied research in one of the world's leading research institutes. We asked him if 3D printing technology was now in a position to fulfill our every wish and desire at any given location?

"Design freedom is enormous in the field of 3D printing. Manufacturing can now be done in an entirely new way: Not in factories in the Far East, for instance, where large quantities of goods are manufactured and then shipped across the globe by container. Instead, manufacturing can be done where the product is needed. This technology enjoys great potential to better distribute complex products. In many fields of medical technology, this has already become the new norm. It’s still not enough though to just plop a printer down in a far-flung place and to press ‘print’ in order to produce a highly complex item, such as an implant. 3D printing requires craftsmanship, engineering skills and expertise. We need to bear in mind that a 3D printer is not some kind of ‘black box’ that can magically ‘conjure’ the wares we need. Because in the end, there are highly complex physical processes taking place inside a printer that require ongoing development and expertise. So it's a little different than on the starship Enterprise. But can these types of processes potentially improve people’s living conditions? Yes, I think so, but we still have some hurdles to contend with before that. Research is driving various issues right now, such as refining resource efficiency with regard to 3D printing or the customizing of specific merchandise. Currently, we are also working on innovations such as smart materials, which enable shape alterations through external influences. 3D printing is a key technology for these programmable materials. A lot will happen in this area in the next few years."

How does autonomous driving work? Will we all have access to our very own personal K.I.T.T., the talking car from the "Knight Rider" series? Dr. Patrick Gebhard is head of the Affective Computing group at the German Research Centre for Artificial Intelligence (DFKI). He researches methods for making communication with technical systems more sociable.

Dr. Gebhard, do cars know what people are thinking?

"In order for an autonomous vehicle to be able to simulate emotions (in the sense of subjective experience) and interact accordingly, the use of affective computing is needed. Affective computing refers to emotion-based AI methods, which supplement computer systems with emotions and social behaviour. This will give rise to increasingly natural interactions with assistance systems. Insights into the human psyche play an important role here: What is the nature of the culture a person grew up in, how has one been socialized, what role do a person’s individual values and norms play? Interpreting emotional patterns in a person’s face or speech do not on their own adequately enable AI to react empathetically. Moreover, we must keep in mind that humans for the most part regulate the externalized expression of their emotions. If the methodology underpinning AI focuses strictly on regulated human expression, the results can be misleading. This is why we’ve chosen to focus on incorporating aspects of regulated human expression into AI computer systems. So to answer your question, an autonomous vehicle that operates without this kind of methodology cannot ‘know’ what we are thinking, it merely sees what we reveal through our facial expressions. There is a big difference here. Currently we’re trying to apply sociological and psychological theories to interactive computer models; in this way, future-oriented autonomous systems will enable us to reap great benefits in fields like mobility or medicine. However, for these systems to serve their purpose, we must be able to trust them. We are studying how this trust might be generated and how the robot, in this case a vehicle, should act. Is it important for an assistance system to have a face? Should it greet the driver? Success also depends on the individual's ability to bond. We mustn’t forget though that AI humanization methodologies find their roots in technological approaches. This methodology cannot create a ‘being.’ At best, it can simulate experience and empathy, but it cannot truly experience anything.."