“AI is about more than just increasing efficiency”
Daniel Abbou, Managing Director of the German AI Association, on AI and the common good, a new culture of sharing data, and the importance of data science as a school subject.
Mr Abbou, the German AI Association, is supporting the Civic Innovation Platform. Why is this project of interest to your organisation?
When we think of Artificial Intelligence, one of the first things that come to mind is raising efficiency. Yet, AI can lead to achieving so much more than this. It can be used in many areas such as applications for the benefit of the common good. The Civic Innovation Platform helps us to reach groups of people for whom AI is currently not yet an issue. We can show them how AI can help to improve society. That’s what makes the platform interesting for us.
As an association, you represent the interests of AI businesses. How is the industry doing at the moment?
If truth be told, we’re doing great. Obviously, the pandemic is leaving its mark on individual industries such as the event and culture sectors, which are going through difficult economic times at the moment. Even so, we were able to gain 80 to 100 new companies as members last year alone. Young companies and AI are currently experiencing a major upswing.
What do you think the reasons are for this upswing despite the coronavirus crisis?
Many digital and AI projects which had previously been postponed by companies have gained higher priority in the current situation. This is clearly reflected in our market.
"Among other things, an AI application for the benefit of the common good addresses people whose only access to the digital world is via their mobile phones. This also includes people with disabilities. If it is able to support individual people who are not privileged, AI will gain enormous significance for society as a whole."
What is your take on the promotion of AI in Germany? What aspects are working well? And in what areas do you see room for improvement?
As I see it, the guidelines for promoting young companies and start-ups in Germany are not always in sync with their reality. The high barriers for obtaining financial support and the various documentation obligations are a problem. Frequently, for example, young companies are unable to plan far into the future, even though this is precisely what is expected of them. The Civic Innovation Platform is a positive example in this respect as the hurdles for participating in it are low, meaning that young companies and start-ups can join it without any problems.
What do you think are the characteristics of a competent AI application for the benefit of the common good and what aspects are decisive in developing such applications?
Among other things, an AI application for the benefit of the common good addresses people whose only access to the digital world is via their mobile phones. This also includes people with disabilities. If it is able to support individual people who are not privileged, AI will gain enormous significance for society as a whole. This concerns very practical aspects such as using AI to reduce frustration at work. Future AI projects will also need to keep an eye on these groups.
Against this backdrop, how did you experience the first round of the “AI is what we make it” idea contest and what would you like to see for the future rounds?
What I liked was the range of ideas, such as an AI box which familiarises children with learning systems in a playful manner, AI-based data evaluation for identifying individual health risks and taking corresponding preventive measures, and ideas for physical exercises at work including real-time feedback on the execution of the exercise. I was also very impressed by how well the project groups and the Federal Ministry of Labour and Social Affairs (BMAS) worked together. It was thrilling to see the Policy Lab embark on an experiment in the form of the platform and the idea contest. As with any experiment, you learn from the experience gained. So, I’m pleased that things are continuing.
Many companies have unused data resources – what can be done to foster a data-sharing culture?
You have to break free from old cultures to achieve a data-sharing culture. Frequently, companies are stuck in old rigid structures. They create data silos and install firewalls to block out the outside world to protect data from plagiarists. And now they’re suddenly expected to share their data in the cloud? This cultural shift is difficult. To set it in motion, we need secure data platforms so that all parties will accept data-sharing and realise that they will not suffer any disadvantages as a result.
You advocate broad-based AI education and training, for example data science as a compulsory subject from the third year of schooling. Why is it important for everyone to learn about AI from an early stage?
I’m not demanding that all schoolchildren be required to learn a programming language. It’s more about basic concepts of data science such as algorithms, data-sharing and if-then loops, in other words, greater digital literacy. Things cannot continue as they are at the moment. In some cases, young people complete their schooling without ever having had any exposure to such concepts. However, they are confronted by this subject when they enter the working world, if not before. So, they need to be digitally literate and know what’s what.
You recently tabled a position paper on the relationship between AI and climate protection. In what ways can AI contribute to greater sustainability?
We have the chance to develop AI products that carry climate protection and sustainability deep in the DNA of their algorithms. These products have potential in the world market. Environmental protection must be seen as an essential part of AI. There’s an apt quote from Luciano Floridi, Professor of Philosophy and Ethics of Information at the University of Oxford, who said, “Sustainability is not the cherry on the cake – it’s the cake”.
Are any AI products already making a decisive contribution to climate protection?
There are some specific applications. Yet, the potential is far from exhausted. On the contrary, it is enormous. One of the key tasks for AI is to detect patterns that are not yet evident to the human eye and brain. A good example of this is the data collected by German institutions or the state surveying technology offices, which have been holding data resources for aeons. If this data were to be digitalised and made accessible, it could be used for detecting patterns and for creating corresponding AI applications. These, in turn, could be used to draw conclusions and to generate forecasts to facilitate sustainable urban planning, for example. Another example is forestry, where drones fitted with cameras are already being used. The resulting images could then be utilised by an AI application to show where woodland needs support due to climate change, e.g. by reforesting with other tree species. Then there’s the construction industry, an important sector in which sustainability does not yet play a key role. I also see great potential here for optimising supply chains and for performing analyses of construction materials. There are many options.
AI is undergoing very dynamic development. If you venture a gaze into the future, what issues do you think the German AI Association will be occupied with 20 years from now?
I don’t think it’s possible to predict what things will be like in 20 years’ time. Looking ahead over the next five years, I would hazard a guess that quantum computing will have an impact on the AI scene. This technology substantially expands the limits of computing efficiency. However, the last few years in particular have shown that there’s so much going on that reliable predictions are impossible. What is certain is that the situation will remain a fascinating one!
Mr Abbou, many thanks for talking to us!