Dr. Andreas Liebl works at Europe’s largest innovation centre UnternehmerTUM, where he set up the appliedAI initiative after observing a gap in terms of support for startups and the AI industry in Europe. “We noticed that there were so many questions out there that lots of people had. We thought that we could try and answer all these questions together so that people wouldn’t be forced to answer these questions individually, over and over.” Questions like how do I get from POC (Proof Of Concept) to production? What sort of knowledge does my existing workforce need to have in the future? What kind of hardware or software infrastructure should I use, which of these are also tools which are allowed in Europe and which align with GDPR?
The initiative’s slogan is “We lift Germany to the AI age.” On one hand, things look increasingly positive for AI in Germany. appliedAI released a report recently which stated that there are now 62% more German startups using AI in their products and services and that since 2009, there’s been around 1.2 billion euros invested in AI startups in Germany. However, this number looks a little less dazzling when we examine the broader context – just one AI startup in China, SenseTime, has had 2.2 billion euros invested into it in the last two years. So is Germany – and more broadly, Europe – losing the AI race?
Liebl emphasizes that the situation is complex. He stresses that since SenseTime focuses on surveillance, there’s governmental interest in it and the startup has contracts with different cities in China to monitor the public. He points out that large players in China “cannot easily move money outside of China due to government restriction on larger sums of money leaving the country, so their money stays within China – as such, they invest it in Chinese startups. If there are startups which have the government’s blessing and which have won huge projects from public institutions then that’s obviously a bonus for the investors, as well.” He believes China’s advantage is that they can “strategically grow its own innovative ecosystem through public projects which are awarded to these types of startups.” He also notes that in China (as in America), much more public domain data is available than in Europe, due to Europe’s privacy regulations, which give startups and the innovation ecosystem outside of Europe a clear advantage.
However, it’s not all gloomy news. There are “quite a few companies which are leading [the push for AI] together with American and Chinese companies, Siemens for example, or Telekom or Allianz. Also, some specialized startups like Artisense or DeepL are leading things from a global perspective.” Plus, he believes that if you look at the B2B field of AI applications, “Europe is leading the world in terms of expert knowledge. Within this field, you really need the knowledge to train the networks…so if we get traction now and start focusing on AI transformation, we could be competitors in terms of B2B.” He is aware that Chinese companies are interested in obtaining this same knowledge from Europe to start using AI industrially. “But if we retain this same knowledge strategically in Europe and if we get going, I believe that we could at least be competitive on the B2B side which has always been a strength for us in Europe.”
On the other hand, he notes, the basis for Germany’s industrial success is “the small or medium-sized enterprise. These are the same companies which are really struggling to find talent, in getting started on working on AI projects,” though he stresses that he doesn’t think this situation is unique to Germany. Liebl also argues that the funding situation isn’t ideal – while funding is growing increasingly globalized and German startups are grateful for money flowing into the local ecosystem from the US and China, he personally “would very much appreciate it if there was more money coming in from Europe itself but [is] aware that there needs to be a push from regulators, from politics to really boost the investments from within Europe.”
Ultimately, Liebl wants clearer direction and support from the German government. He alludes to the push the Chinese government has given to AI and how we can see this in both funding and innovation in China. “I would say the most important thing is that we need to have a clear vision of what Germany wants to do with this new technology and to get everyone aligned on these common goals. If the government would say tomorrow that Germany should aim to be leaders in healthcare and AI; that would create a push for insurance companies [to get on board], for required regulatory changes, and would set priorities to get things implemented. Most importantly, this would push all stakeholders to work aligned and in parallel on the same targets. Because we don’t have a really specific vision of what we want to pursue and because we don’t have an alignment of the stakeholders,that really slows us down.”
But there are still AI projects in Europe to get excited about. He namechecks AI HLEG, a high level expert group on AI for the European Commission, which provides guidance in the European Union about the subject. He applauds AI HLEG for the emphasis they place on building trustworthy AI. Recently, they have published ethical guidelines providing recommendations for accomplishing this, something he feels “really provides a narrative as to what we want to accomplish as Europe.”
There’s a lot to think about. There’s obviously so much potential, he points out, in terms of how AI can affect every aspect of our lives, both privately and professionally. The question that AI raises for him is “how we want to live in the future. We live a bit differently in Europe than in China or the US with the value system we have.” He believes if we want to continue to have this value system and also use this new technology, Europe will be forced to relinquish the observer role it currently holds and become more active, “because what we’re building now will really influence what happens in our lives over the next few decades.”
“If you talk to public bodies what they say is that finding a consensus at the moment, given our democratic process, takes time.” The problem is that this means that decisions are taking place more slowly than the technology itself is being developed. “So we’re always running behind because the nature of our system, which is based on consensus, takes time. This means we have a systematic disadvantage at the moment because the development is taking place so fast.” He stops and thinks. “We need to find a way to make the consensus system a strength once again through resilience, variety, diversity.”