“AI can be an even bigger opportunity than the internet boom of the late 90s”

Mr. Neubert, your LinkedIn profile describes you as an “evangelist for Transatlantic AI eXchange.” Is artificial intelligence (AI) a kind of creed to you?

Neubert: (laughs) No, it’s not. It’s funny. My wife actually asked the same question when I printed that on my business card. Intel categorized 16 vertical applications where AI will enhance things and make them better, more efficient, cost effective, faster, more accurate.  It will affect every single aspect of our lives one way or the other. So for me, it is more about scaling horizontally. And therefore, it’s an equally great or even bigger opportunity than the internet in the late 90s. The development might not be that drastic, but it is going to change the world.

According to Kate Crawford, Microsoft researcher and author of the “Atlas of AI,” artificial intelligence is “neither artificial nor intelligent,” but rather created from natural materials, with humans performing the tasks that make the AI appear autonomous.

Neubert: I would only halfway agree with that statement. AI needs to have a front-end tool or a set of data that humans have collected, so it needs the human input. That’s true. But people need to understand that deep learning and artificial intelligence is a subset of machine learning, not the other way around. For me, the uniqueness of AI is that the data can be fed back into the engine, and it learns from its mistakes to be better, faster, and more accurate. In other words: It can correct itself.

A relatively homogeneous group of professionals in an affluent industry is currently driving every major technological breakthrough in AI. This may sound provocative, but isn’t this a sort of ivory tower?

Neubert: Think of the famous bell curve of development.  With AI, we are still in the early adopter’s stage.  The US has taken the lead because of visionary corporations that bought deep tech companies at some point. China is very aggressive and is also doing much more than anyone else. So, the gap is between these two big countries who have these giant leading high-tech companies, and the rest of the world. The tools that are being used to develop AI applications are still very clumsy. A certain level of education is needed to use them. So at the moment, it’s really a race as to who can develop the best AI foundations at a global scale. And in that regard, I would say everybody is working on this, in one way or the other.

Many AI campuses are currently popping up in Germany. Are these developments similar to what we once saw in Silicon Valley?

Neubert: There are a few key cities with the most established technical universities: Berlin, Hamburg, Munich. I would also name Karlsruhe, Aachen, and Tübingen. They might also catch up with the AI development bandwagon. Saarbrücken is home to the DFKI, the German Research Center for Artificial Intelligence (DFKI). The newly emerging AI campuses are affiliated with these universities and facilities. They may even go a step further and start taking money from private investors, similarly to how it is being done in Silicon Valley. They need to start very early on to build accelerator and incubator programs to establish a direct link to the university curriculum and get students on track to merge right into startup-entrepreneurship. That approach is not very well developed in Germany. It’s a cultural, structural thing.  It’s breaking right now, which is positive. And then, in combination with these research centers, industry must come in: car and medical device manufacturers, healthcare biotechnology manufacturers. They will need to collaborate closely with these centers to develop the next generation of AI products. It is about setting up prototypes of use cases, and the industry taking them, customizing them, and sending them to the market.

A year ago, you launched the Transatlantic AI eXchange Platform. How can it help drive these developments?

Neubert: Right now, it is a platform for webinar series and workshops. It brings together influencers from industry, research, and government to examine trends and moonshot AI use cases.  As of December 2021, we have had seven sessions.  Another twelve are coming in 2022. We would like to ignite the transatlantic AI exchange. We also want to learn from the Europeans how the US can better localize its AI efforts to meet certain requirements that apply in the EU.

Is the German business ecosystem ready for the advance of AI?

Neubert: I think Germany has acknowledged that it needs to stay competitive and not miss out on a fourth mega shift within the industry in the past 20 years.  Is the startup community ready? Is the younger generation willing to be more entrepreneurial? In my opinion, the answer is absolutely yes. Is the whole ecosystem ready? No, because there are bigger corporations. There’s the middle management problem. It is the current generation of leadership executives. We need to bring these two worlds together. We need agility and the right mindset, to allow people to fail and try again. That’s what makes Silicon Valley so beautiful.  Big German corporations have the money, the potential to adopt AI as quickly as they can to enhance their existing business or shift into new businesses. Some of them are already making heavy investments within their own AI departments. They hire external specialists, AI professors from universities. As a next step, they will be building their own internal knowhow, and then pretty much drive themselves.

What impact will the European Data Protection Regulation and data visibility regulations have on the industry?

Neubert: Let the German government and Brussels start working on supporting regulations because it will take a long time. In the meantime, Berlin can catch up on developing technologies. If American or Chinese companies want to sell their services based on their AI engines in Europe, they will need to meet European regulations, to acknowledge and establish the same kind of standards in their products, so that they can keep selling their products to European customers. So I am all for it. AI is too sensitive. Too many lives can be affected by wrong decisions. We need to step back, talk about the danger, articulate and address it, and then go from there. We also need visibility of AI models and graphs so that people actually understand what is happening within an AI engine. Why did it make that particular decision? Because if we don’t know, it can have a very negative effect.

Could it be that by the time Europe introduces these complex regulations, technology will already be one step ahead? Will legislation lag too far behind?

Neubert: Yes, there’s a little bit of a danger there. And the challenge is twofold.  European regulatory bodies must be more effective, react faster, and be more efficient. If they don’t acknowledge that, we have a problem. The first regulatory baseline needs to be established as fast as possible, a basic framework that generally addresses the sensitive aspects of artificial intelligence. And then, that baseline needs to be revised to catch up with new development cycles.  And that’s probably an ongoing, continuous loop.  This is normal, because innovation is usually one or two steps ahead of regulation.

As chairman of the German American Business Association, have you felt any new impetus from the Biden administration?

Neubert: It’s certainly a drastic contrast to the previous administration.  We also have a new government in Berlin. I believe that its mindset on AI is very similar to that of the Biden administration. So I think this is going to be a very synergetic relationship. We are at a new beginning in redefining and rebuilding that relationship of trust on the political level. On the practical level, I do believe that both countries have acknowledged that they actually need each other. The U.S. is far ahead on AI development and integration of AI in vertical applications and Germany is trying to do the best it can to catch up. That’s not really a competition.  At the same time, the U.S. has recognized that Europe is probably one of the biggest markets to sell its AI products – apart from China with all its difficulties. So there is a joint interest to help each other.

You have been living in Silicon Valley for 30 years. Will it have to reinvent itself in the light of new breakthroughs in AI?

Neubert: When I came to Silicon Valley, there was no mobile phone. There was no videoconferencing. I witnessed both the beginnings of the internet hype and the emergence of its downsides, until the bubble burst. It was bitter, it was hard, but we came out much stronger than before. And then the financial crisis hit. And that was bad, too. The Valley has seen a transformation from semiconductor- and hardware-centric to internet- and then to software-centric.

Now I think we are right at the start of a phase where everything – well, not everything, but a lot – will revolve around AI and its implementation.  Will it be similar to the crazy times when the internet was developed? No, not at all. Because everybody learned their lesson. And this is not something entirely novel, because AI is more of an enhancement.  Also, similar developments will be happening around the globe. The technology will democratize extremely quickly. The infrastructure, the software, public cloud services, affordable hardware, are already there. So, the AI phase will go global much faster and less painfully than the internet phase.

If you had the opportunity to develop an AI-based application as a startup founder right now, what would it be?

Neubert: Well, if I were a smart entrepreneur, I wouldn’t tell you (laughs).  But jokes aside, unfortunately, I don’t have a brilliant idea yet. Or I may not have run into the right startup company yet.  Education is the field where I would most like to leverage artificial intelligence to make a positive change in the world. So, if somebody has a startup company or an idea in mind that could get me fired up…

A German version of this interview was first published by BigData-Insider on January 10, 2022. Research for this article was made possible with the support of the Heinrich Boell Foundation Washington, DC.