“When it comes to using AI for the design of infrastructure (roads, cities, etc.), one of the most important challenges to overcome is the creation of detailed Digital Twin simulations.”

20 years old, Canadian, and striving to develop the future of technology: Shalev Lifshitz is one of the world’s youngest AI researchers and entrepreneurs. He has spoken at various international conferences, inspiring many to think about the future of AI and to explore which questions need to be asked to prepare for human-level artificial intelligence. Shalev is creating AI algorithms to discover optimal actions in 3D digital twins of reality. He is building algorithms to detect the causes of crashes in cities and conducts Deep Reinforcement Learning research.

In our interview, the up-and-coming young researcher talks about his research areas of artificial intelligence, deep learning and digital twins – and, of course, their areas of application in the mobility transition. Furthermore, he talks about the use of AI methods in the design of sustainable urban environments and how digital twin simulations can be used meaningfully.

What will happen if we don´t shift our mobility behaviour?

SL: For myself, shifting our mobility means implementing and integrating new applications of technology into our existing transportation systems, or creating completely new transportation systems in general. There will be multiple effects if we stagnate and do not make such shifts. Of course, the first result is that non-renewable energy sources (particularly fossil fuels) will continue to power the industry, leading to further global warming and increasingly frequent climate catastrophes. The second result of not innovating is that we won’t be able to save those who would otherwise lose their lives in mobility related accidents. This result can be avoided by utilizing technologies such as AI, IoT, and 3D Digital Twins. In my talk, I will discuss how we can use these technologies to avoid such a result.

What do we have to radically invent, improve or change to realize the turnaround in transport policy?

SL: Let’s consider our switch away from fossil fuels to cleaner and renewable energy sources. Before we can talk about how policy can help, we must talk about a major technological obstacle: the current power grid. The power grid must be modified and updated, so that homes can feed solar-generated power back into the grid itself. In this way, we allow for more distributed energy creation. Once this is made widely available, we then need to update policy to incentivize homeowners to do so. Of course, we also need to financially incentivize large companies to move to clean, renewable energy sources.

What is the most hyped buzzword in terms of mobility which has in your opinion no impact on the real issues of mobility?

SL: I’ve recently heard about the idea of the ‘gamification’ of cities, which from my understanding is meant to attract younger audiences to interact with cities through technology-based games. Although I think that such an idea could be a fun addition to cities, I still believe that most of our effort should be focused on solving more important mobility issues, such as moving away from fossil fuels and eliminating mobility-related accidents.

What will be the most important thing in 10 years associated with “mobility” that comes to your mind?

SL: The ability of AI agents to design optimal cities in detailed simulations of reality.

If you were able to use only ONE mobility solution for the rest of your life – what vehicle/ mobility solution would you choose?

SL: Although I would really hate to give up my forest bike rides, I would probably go with electric cars since they provide the best balance between short and long-range trips. Also, we will hopefully get full-out autonomy soon, which would add another level of convenience (though it might yield other challenges in terms of road congestion).

What is the greatest mobility challenge for your research field these days?

SL: When it comes to using AI for the design of infrastructure (roads, cities, etc.), one of the most important challenges to overcome is the creation of detailed digital twin simulations. For the AI’s discoveries in simulation to be useful in reality, these simulations need to be detailed and realistic – which is hard to achieve. Work needs to be done to enable faster creation of these realistic simulations.