Interview on “Artificial Intelligence (AI) with Impact” with the Web Summit 2020 Masterclass speakers.
At this year’s Web Summit, Jenny Bofinger-Schuster, Head of Sustainability and Benno Blumoser, Innovation Head of the Siemens AI Lab, described in their Masterclass “AI with Impact: Can AI create a more sustainable world?” where Artificial Intelligence can help us to solve some of our major challenges. We talked to Jenny and Benno to share some key takeaways from the session and which developments they see on the horizon.
Question: You talked about “AI with impact” on the Web Summit, which impact do you mean?
Jenny: When I think of impact, I think of course of sustainability and the environmental impact, e.g. reducing emissions and reducing waste as well as of social impact like the sustainable development of societies. Megatrends have a strong relevance for our future and with AI, we can shape their impact for the better. Let’s have a look at some questions in this context:
- The Future of Mobility: How can we establish a connected, shared and electrified mobility system?
- The Future of Cities: About 80% of emissions in the future will be produced in cities, so how can we improve the overall quality of life for all there?
- The Future of Work: How can work be more inclusive and healthier?
- The Future of Healthcare: How can it be more personalized and more available to the people around the world? And how can we connect the technology with the human expertise that is there?
- The Future of Energy: How can we shape the future of energy, so it is more decarbonized and decentralized?
- Industry 4.0: How can we have the industry of the future be more data driven, flexible and finally sustainable?
Benno: Well, there are a lot of “futures” in there. Artificial Intelligence is a powerful tool that can in fact contribute a lot to answer these questions, be it through AI-driven prediction capabilities that lead to improved resource and energy efficiency, through visual intelligence that allow to develop autonomous and “zero harm” vehicles of any kind, or through augmenting the jobs that anyway make the world a better place, if you think of the health or the education sector.
Question: And where can we see this impact in our everyday reality?
Jenny: One “future” that relates to all of us is mobility. Generally, public transport in trains and buses is core for sustainable urban mobility – taking little space per traveler, producing little or no emission – etc. But then, also in the inner city, there is the first and last mile, that need to be covered, by foot, by bike – or in a shuttle!
We believe that huge impact on sustainable urban mobility can be achieved through demand-responsive shared autonomous mobility solutions. Driverless mobility is not a new thing to Siemens when it comes to subways or also trams. Still, in road traffic autonomous driving is a challenge with multiple road users active on site and multiple intersections along the way!
We see the infrastructure on the road as a decisive element when it comes to truly sustainable safe solutions. In October, Siemens Mobility started one of the first pilots for autonomous driving in Hamburg. The project is called HEAT for Hamburg Electric Autonomous Transportation – and it is one of the first projects worldwide where autonomous cars are running in a high urban density area.
AI plays a key role when it comes to the further development of autonomous driving:
AI is applied when it comes to understanding the prediction of the behavior of what we call the “vulnerable road user”, i.e. especially the pedestrians.
We can detect patterns of accidents and identify high risk locations – and thereby enable the improvement of driving in the city – a very important asset even before introducing autonomous cars. This facilitates positive public opinion and is also a real asset for city governments.
Finally, with public shuttles w/o a driver, safety includes other components as well: AI can be used for aggression detection and passenger condition monitoring, or we can detect stressful or critical situations for the passengers and raise the alarm once life & limb is at risk. You see, AI and multi-sensor systems can improve safety inside the public shuttles as well.
Question: Autonomous cars is always a great example to show what AI can do with our lives. But how can our customers in other areas, for example in the industry, benefit from Artificial Intelligence?
Benno: Unfolding the potential of AI in the industry requires both a deep understanding of processes and the expert knowledge on AI and related technologies. This is why the application of AI is always a true journey that we go together with our customers.
If I look at Industrial AI from a sustainability perspective, handling of waste is one of the key challenges because it is at the intersection of two important sustainability levers: automating the process makes it more resource-efficient and improving the recycling process in particular focuses on the cornerstone of the circular economy. So, one of our customers had an issue since the existing algorithm for garbage sorting was mainly based on a simple machine learning model with problems in the operation and low accuracy rates when it came to automatically detecting the different types of industrial waste or interpreting new emerging waste types.
So, our Siemens colleagues solved the quality issues that emerged from the unbalanced data sets by applying transfer learning, one-shot learning and other techniques that helped to improve the image classification accuracy at the fringes of the data set. In the end, they achieved an accuracy rate of more than 97% for the various classes of waste, and a came up with an approach to detect new type of waste in time. Another insightful part of the story is, that during the project, the algorithm was also optimized by active learning routines, meaning that humans validated or corrected the machine sorting behavior and thus help it to improve over time. This is a very nice example for human-machine team work which eventually delivers the best (and most sustainable) results.
Question: What was your most important messages to address to the audience at Web Summit?
Benno: AI is here to add important value to solve well-known and already existing, everyday challenges. You need open-minded people contributing with their respective domain knowledge, technical expertise and problem-solving skills to come up with a robust solution. This only works by collaborating across functions and company borders. And we have to learn to explore – every use case provides us with new insights, lessons learned or even best practices. From a business side, it’s important to scale working solutions, but it’s equally essential to stay curious to find new approaches and use cases that may even better address the key challenges of the future.
Jenny: I’m convinced that the future is digital AND sustainable. And we clearly want to shape this future in the fields we are active in. How do we aspire the future to look like and what does it need to get there? I do believe that AI plays a very important role. If applied in the right way it will help to improve the quality of life and increase efficiency in all areas. It can transform so many areas which are essential for our everyday lives, be it with respect to energy, water, material – e.g. by reducing carbon emissions or foster circular economy and much more. You need to understand where your biggest impact is – in good and in bad. Then you have to get started and find solutions. Most importantly for the companies out there: get started, make the first step – and you may end big!