Euro-CASE Conference 2018: The morning
Artificial Intelligence was the focus topic of this year’s Euro-CASE Conference held on 24 September in Rüschlikon near Zurich. SATW, a member academy of Euro-CASE, organised the event.
The Forum at the Swiss Re Centre for Global Dialogue was well filled when Willy R. Gehrer, president of SATW, welcomed all the participants, in particular Government Councillor Carmen Walker Späh and State Secretary Mauro Dell'Ambrogio. In his short introduction, Willy R. Gehrer pointed out why Artificial Intelligence (AI) was chosen as the topic. “AI is simply THE current tech topic. And it goes far beyond technology. AI has the potential to fundamentally change society.”
To go further than just presenting strategies
To conclude, Willy R. Gehrer introduced Ana Maria Montero of CNN Money Switzerland, the moderator of the day. She took over and presented the next speaker: Reinhard Hüttl, president of Euro-CASE, briefly introduced the organisation. Euro-CASE has 23 member academies from all over Europe. One of its main goals is to provide evidence-based and balanced expertise. Among other things, Euro-CASE is member of SAPEA, which unites over 100 academies from 40 European countries. Under this umbrella, societal challenges can be tackled in an interdisciplinary manner. Reinhard Hüttl announced that a European strategy on AI is to be published before the end of this year. “But today we will go further than just presenting strategies. The future is ours to shape.”
Then followed the welcome by the representative of the canton Zurich, Government Councillor Carmen Walker Späh. She mentioned the many ways AI can be used, e.g. autonomous cars or facial recognition. Zurich has an excellent reputation for ICT research with companies like Google or IBM and universities like ETH Zurich. The canton wants to stay an attractive location for ICT companies and all these clever minds. But with such developments as AI, the state must always be at the forefront when it comes to anticipating the social consequences.
Jeffrey Bohn concluded the welcoming round on behalf of the host – the Swiss Re Institute – presenting their research in the field of Machine Learning. Swiss Re has been doing research for a long time and in a broad field with teams all over the world. Open Data is very important for their research because science should be “show me” not “trust me”. Jeffrey Bohn pointed out that the Swiss Re Institute is committed to raising the bar for scientific research. “There are so many terms besides AI that exist. Some frighten people. Others are very useful. We should talk more about them as well.” He pointed out that the focus should be more on addressing the challenge of data curation as well as testing algorithms on real useful data on scale. Furthermore, it should be discussed how AI integrates into a digitising society.
Innovation policy in times of digitalisation
Mauro Dell'Ambrogio, head of the state secretariat of education, research and innovation, gave the first keynote about innovation policy in general and especially in times of digitalisation. “Because we don't know what is coming preparing for the future is always difficult. Either we prepare for the most likely or we create a flexible environment that can then be adapted. That's what we do in Switzerland.” So far, innovation in Switzerland has worked well. Together with the cantons, the federal government is promoting ICT skills at schools and in continuing education. And the next national research programmes will be on digital transformation. NRP77 has already been announced; more will follow by the end of the year. In Switzerland, already many researchers and companies are in the driving seat. Mauro Dell’Ambrogio pointed out that it is the duty of the government to support them.
“We need to get AI right”The second keynote was delivered by Alessandro Curioni, director of IBM Research – Zurich. “AI has positive effect on business and on society but we need to get it right.” Firstly, he spoke about the research at IBM and secondly about a national AI strategy for Switzerland. Why is AI so interesting today? “Finally, we have enough data, we have enough good algorithms, but above all we have enough computer performance.” But most data are not used. And a lot of them quickly lose their value. AI can help to optimise decisions across all industries form finance to science, e.g. in predicting chemical reactions. In the evolution of AI, at some point General AI will be reached. “But way before that we will reach Broad AI which is disruptive, pervasive and hence important for business and society.” Alessandro Curioni made a plea for transparency and trust. Only companies that can achieve this will be successful in the future. “The problem is that we base AI on data from the past with all the biases.” Technologies must be developed that recognise and correct these biases. This is the only way to achieve trusted AI. The biggest problem we face today is trust. IBM research is very engaged in trusted AI.
The advantage is that Switzerland has a traditional reputation as a trusted third party. Switzerland could create an open market for data and empower citizens to become aggregators, controllers and beneficiaries of their own personal data. Alessandro Curioni pointed out the ideas of a certification body for AI systems and a regulatory framework. AI systems should be explainable, ethical, bias-free. As most companies in Switzerland are SME it is also very important to enable them with ease of use and automation and for instance “deep learning as a service”.
Address the right challenges
Professor Luciano Floridi from Oxford University, the last keynote of the morning, talked about ethics and AI. “If anyone asks if you are online, then he or she is from the 80s because today we are “onlife”. Presence is no longer equal to location and agency is not intelligence.” Agency is simply the capacity to perform a task to fulfil a goal. He thinks that the world is getting better with AI but challenges must be met. “It’s not Terminator and all these crazy stories which do only distract.” Insofar as AI is fuelled by data, AI inherits some of its ethical challenges form the debate on data governance, especially consent, ownership, and privacy. He said that AI could be used – these are the opportunities – or abused – these are the risks – but also underused! AI should enhance human decision/control and support human responsibility. “But we should be careful what decisions we delegate to machines.” AI will make human skills needless, but not jobs. There are many skills we like to give up. “So, the mowing robot won’t replace the gardener because he can finally take care of the roses.” From Luciano Floridi’s point of view the new challenge is not technological innovation but the governance of the digital. “We should think deeper, be mindful, care more, and design better. We are not in a hurry. We should take the time to do it right.” Because AI should not only be acceptable but preferable by the society.
For the first panel "How AI Can Take off in Europe" Professor Jana Koehler, Lucerne University of Applied Sciences and Arts, and Isabelle Kovacsovics, Ringier, joined the three keynote speakers. China and other regions of the world are in the lead. How far back is Europe, asked Ana Maria Montero? According to Isabelle Kovacsovics, the media industry is 3 to 5 years behind the USA and China. But we are catching up. Alessandro Curioni doesn’t mind to be a little later, if things are done better. “Speed is not everything. We have to do it right.” “What can government do?” was the question addressed to Mauro Dell'Ambrogio: “We can create a good environment and cooperate with European partners.” For Jana Koehler, AI research in Europe is not lagging behind. The difference is in the funding. Researcher in Europe have to spend too much time writing proposals. Still, Europe has more researchers in AI than any other region, as Luciano Floridi pointed out. “For instance, robotics is at home in Europe.” But Europe should do more research together. “We have good companies, but they are all taken over by American or Asian companies”, said Jana Koehler.
So, does a country need an AI strategy? Luciano Floridi highlighted an important point: A national strategy is not for the nation alone but helps them to talk with each other. AI pushes ethics to their limits. “It is a wonderful time for philosophers,” said Luciano Floridi. “So how about Ethics as a service?” Alessandro Curioni: Everything starts with the data. Open Data is good, but we should do more.” I can donate my body for research, but not my medical data”, added Luciano Floridi. New regulation is needed. Personally affected, Ana Maria Montero wanted to know what AI means for the media industry? Isabelle Kovacsovics had to admit that although there are many good people in Switzerland, it is difficult to win them over to work for a media company because most prefer to join Google. “Information is everywhere but not quality”, said Jana Kohler. “How do we bring quality journalism back?” This is a major challenge for our democratic society. Maybe AI can help? The open discussion addressed the subject of big data. Other countries and companies have much more date, so can Europe and European companies successfully compete? Alessandro Curioni pointed out that the quality of data is more important than its mass. He believes, there is enough data in Europe. Jana Koehler added that AI can even help to bring the industry back to Europe.