Rise of AI Conference 2019 · Berlin

AI Is Not a Threat, but a Chance for Our Society

Andreas Steier · Member of the German Bundestag

At Rise of AI 2019 in Berlin, Andreas Steier, Member of the German Bundestag, argues that AI policy in Germany should build on existing technical standards rather than start from fear. A former automotive engineer who worked on sensor systems and neural networks before 'AI' became a marketing term, he describes joining the Bundestag in 2017 and becoming his party group's rapporteur for artificial intelligence, sitting on a Bundestag commission that works on an ethical framework for AI. He says the dominant reaction he meets outside Berlin is fear, often based on fiction rather than the reality of the technology, and argues Germany should extend standards already proven in the automotive, energy, industrial and health sectors into the digital age rather than treat AI as an unprecedented threat. Comparing Germany's position to the more entrepreneurial United States and a more centrally directed China, he argues Germany's advantage lies in engineering quality and functional safety, illustrated with a 2018 collision in the United States between a self-driving car and a cyclist, and proposes synthetic data as a way to test rare, safety-critical driving situations.

Key takeaways

  • Steier worked on neural sensor systems for the automotive industry before 'artificial intelligence' became the marketing term for such technology.
  • As his party's Bundestag rapporteur for AI, he sits on a Bundestag commission developing an ethical framework for artificial intelligence.
  • He argues the main obstacle to AI acceptance outside Berlin is fear based on fictional depictions rather than the technology's reality.
  • He proposes extending existing automotive, energy, industrial and health-sector safety standards into digital AI regulation rather than writing rules from scratch.
  • He points to synthetic data as a way German companies can test rare, safety-critical situations for autonomous driving, citing a 2018 self-driving car collision with a cyclist in the United States.

Detailed notes

From automotive engineer to Bundestag AI rapporteur

Steier describes his earlier career as an engineer in the automotive industry, developing sensor systems in Luxembourg and internationally and holding patents on systems that used neural methods to distinguish people from other objects, at a time before 'artificial intelligence' was the term used to market such products. He entered the Bundestag in 2017, became a member of the Council for Research and Education, and serves as his party group's rapporteur for artificial intelligence, a role in which he sits on a Bundestag commission on AI working with technical experts, including former IT specialists now in the Bundestag, to develop an ethical framework.

Addressing public fear: separating fiction from reality

Steier says that when he discusses AI with people outside Berlin, the dominant reaction is fear, with many picturing cyborgs or scenarios drawn from fiction rather than the actual technology. He argues that AI should be treated as a development rather than a revolution, comparable to earlier new inventions, and that the political task is to keep people at the center of decisions while using existing regulatory structures already proven in the automotive, energy, industrial and health sectors, adapting them from analog standards into digital ones.

A German path: quality, synthetic data and functional safety

Comparing national approaches, Steier says the United States has stronger entrepreneurial dynamism for turning new knowledge into consumer products, while China, in his view, has a different ethical commitment in which the individual is more subordinated to the wider society. He argues Germany's competitive path is quality and functional safety, built on its established reputation in engineering and its Industrie 4.0 expertise. Citing the 2018 case of a self-driving car striking a cyclist in the United States, he argues that although most driving situations are not safety-critical, the rare critical ones must be systematically tested, and points to German companies developing synthetic data to model and test such extreme scenarios as part of a federal government master plan for AI and improved knowledge transfer into industry.

Quotes

Artificial intelligence is like any other new invention.

Andreas Steier

We have to distinguish between what is really real and what is fiction.

Andreas Steier

Questions answered in this talk

How does Andreas Steier think Germany should respond to public fear of AI?

According to Andreas Steier at Rise of AI 2019, Germany should help people separate fiction from reality about AI and extend regulatory standards already trusted in the automotive, energy, industrial and health sectors into digital-age AI rules, rather than treating AI as an unprecedented threat.

What role does synthetic data play in Steier's argument for autonomous vehicle safety?

According to Andreas Steier at Rise of AI 2019, synthetic data lets companies test rare, safety-critical driving situations, such as the kind involved in a 2018 collision between a self-driving car and a cyclist in the United States, that occur too infrequently to rely on real-world data alone.

What is Andreas Steier's role on AI policy in the German Bundestag?

According to Andreas Steier at Rise of AI 2019, he is his party group's rapporteur for artificial intelligence and a member of a Bundestag commission on AI that works with technical experts on an ethical framework for the technology.

Mentioned

synthetic data, functional safety, autonomous driving, Industrie 4.0, AI ethical framework, AI master plan

Keep exploring

More from Rise of AI Conference 2019

More on Geopolitics & Sovereignty

Transcript

Auto-generated transcript, not edited. · 1,677 words

Show the full transcript

usually I'm not original politician because I used to work to enemy yes almost as an engineer and we developed in Luxembourg and all over the world sensor systems for the automotive industry and in that position I also developed special sensors I own some patents on that side and we introduced already at that time numeral systems in order to analyze the data and discriminate the different positions of people discriminate between human beings and between other things and in that case I learned already a lot of neuronal systems at that time AI artificial intelligence was not the main key driving word which was now so-called pronounced and is now at that time we just used in order to sell the product with more quality and in that case I came in then in 2017 into the Bundestag and then in that position I'm now member of the Council for research and education and in my position as Rapporteur for our party group I'm responsible for artificial intelligence and there I want to discuss with you some about

some political things which might be interesting for you and which might also be interesting for the whole situation here in Germany and to see what needs to be developed and how we are set up in our political atmosphere if it comes to artificial intelligence and that's a situation which I always receive when I discuss it in the countryside with people from different kind of society there's always a kind of fear which people feel when they think about artificial intelligence and you see that it's not only a chance but a lot of people they think it's more an disadvantage and there's a danger behind artificial intelligence and in order to that also to certain to improve certain tests systems and - yeah - also enhance people that they know what is really artificial intelligence that's our basic topic in the political discussion in Berlin as well as on especially on the countryside and I see especially there is a certain thing between we are what is real and what is fiction and we have to distinguish

between what is really real thing and what is fiction and most of the people that think when it comes to artifice artificial intelligence it comes there that they think okay there's the cyborg already in their mind the loot end data is already reality and that's the thing basic thing of fear where we have to convince people that it's not like that that artificial intelligence has chances and that real things are not like maybe authors are using and describing in in in books and and for that we have to develop certain things and to get into discussion with people we on the political side here introduced a strategy on artificial intelligence and we also introduced on ket Commission for artificial intelligence which is also a discussion want to show also people what is really behind artificial intelligence and they are experts in it some of them are here some colleagues from the Bundestag who also used to work as informatics experts before they came into the Bundestag are involved but there are also experts in this

on head Commission and we are discussing there an ethical framework so that people are more convinced on this topic another point is what I was going to point out is when it comes to reality we also need to convince people by structures which are used on other topics artificial intelligence is like any other new invention there was always discussion about new invention so it's an in in development it's not I don't use the word revolution it's the development of new possibilities which are also providing and supporting people and that's the question how we can use existing structure how we can put a human being how we can put the man in the middle of the decision and how we can then Duvall evolved out of these structures new standards and we are land of really developing standards in the past my background as an automotive industry and in the automotive industry we developed already standards in the past which give also people a safety security we also developed standards in that position where experts are

sitting together and they are defining a framework which also keeps security on it and that's a main topic where we have to think about is how we can then from our legal point of view how we can use these structures which are already existing on the automotive side and existing on the energy side they exist on the industry side there as well existing on the health industry how we can then improve these standards and how we can put them from one from the analog time into the digital ditch at a time and that's basically my point of view in this audience discussion one's not the lawyer the legal construction have to give a certain loss we have to create a certain framework and in detail that has to be then developed and set up by these nominees and by these committees where experts are sitting together and creating trust is also basic saying what how can we create trust and that's a thing what we need also discuss how can we use existing standards how

can we implement these standards into new things and how can we use for especially our knowledge our experience here from our times we have a big knowledge in Germany and the artificial intelligence is not only a new word in Germany we have a quite high community and science and research which developed already there are milestones in in artificial intelligence and if I consider here the day of Korean Germany which is over 30 years old they developed already milestones in industry 4.0 and that's basically one point where we are so yeah delivering to the world our expertise and now it comes to the next step how can we also use our existing industry our middle stand where we are leading a factor a number one in the world how can we use these expertise in order to develop them our strengths and if it comes in a development in comparison to other countries like China or like the United States we have to see that the United States they are more used in the F DNA

of an entrepreneurship in their mind they know how they can also make money out of new knowledge a lot of new inventions and we somewhere behind in this dynamic markets especially if it comes to the consumer market there we have a little bit lake in comparison to the United States and if you see then on the eastern side like China raising star they are having another ethical commitment to that how they can use that how the person has to somehow order themselves under the big society and that's a challenge where we have to face ourselves and how how we can find our way in this challenge and if we come to Germany I would express that there are some things where we can really with the competition and if we concentrate on our expertise especially on quality like in the world when I came into the United States the German engineer is always a big factor and it's well known in the world and that we have to transfer now to the new time of

artificial intelligence that people are trusting in our knowledge and our experience in our development and that would be a big challenge and big chance for us that we can also create a German way of artificial intelligence how a German can really present our products in the world and then it comes also to the question of functional safety we can only convince people in a functional safety if our products are in a certain way better as other products and if people are convinced in our and if it comes there my past experience I develop different systems for the automotive industry and if you see that a car like in the United State last year that that hit a bicycle while I passing the road then we have also the thing about which data we have to use and the way of which data we have to use we have to also think about what is really critical in reality over 90% of the situations are not critical for autonomous driving if I use that example but

the critical things are quiet then don't happen that often so we have to think about how can we really use a system and how can we also create a model which also tests these critical situations and a critical situation can happen like if a biasing of either passes to treat I always know that in the United States they are not many bicycle riders on the road but they're also not many bicycle violence on an expressway and that needs to be also somehow tested and in that way I know that a lot of German companies are also trying to create synthetic data and with this synthetic data you can also prove extreme situation and you can find also standard how you can also test our system and that would be a chance for us that we concentrate on our quality that we develop our competence centers in that way to also prove a system and that we also convince people with quality and that would be our way of international competition and in that way we

are developing from our federal government a topic a master plan and that needs to by now also develop in our system we need to see also that we have acquired of code transfer in our in the industry and that's the question where we can discuss later than in the discussion well thank you very much

Rise of AI 2027 · Berlin · May 2027

Be in the room for the next one.

Ten years of talks started as conversations in one curated room in Berlin. The next edition brings 300 selected leaders together in May 2027.