AI education strategy, a scientist’s opinion
Interview with Holger H. Hoos, Professor of Machine Learning at Universiteit Leiden (NL).
What is CLAIRE?
CLAIRE is a large network of AI researchers, including the European Space Agency and national governments, namely Belgium, the Czech Republic, Italy and Luxembourg, with additional national governments about to join. As of December last year, we have an actual network of AI researchers and institutes across Europe. That is the largest network of its kind in the world with about 14 000 people. That network allows us to actually operate on a long-term basis to facilitate AI research. At the moment, we are connected experts informally working together, developing a common vision and involving others until we become an actual organisation.
When we look at the non-experts, how do you see the general knowledge about AI and especially algorithms in the general public?
The general population, as well as decision-makers, politicians and administrators, are generally not sufficiently educated on AI to take expert decisions. I’ve heard that in the German Bundestag only three people have basic knowledge in computer science. Let’s not even talk about AI, this is another level. So national parliaments and the European institutions are quite dependent on experts. That is not a new situation in politics – experts are consulted for many political decisions. But because AI is very different from other technologies and essentially is about to touch absolutely everything we do as societies and as citizens, I think AI and political decisions are a bit of a special case.
How is that related to national or European-wide investments?
Especially because the investment in AI matters because it is about developing potential in the right way – not just to be globally competitive, not only for profit-oriented companies in a way that might go against European values. I think it is also very important that some of the general misunderstandings regarding AI are being addressed. Finland’s societal education strategy is a very good model – also in light of what I perceive to be the biggest threat associated with AI in the short and midterm: I think people are quite worried about the ‘wrong’ aspects when thinking about AI. They’re worried about some Hollywood-desk notion of AI taking over, right? Or that AI will take all the jobs within a few years. Actually, both scenarios are extremely unlikely to happen in the near future.
But what are real-world concerns, maybe also from scientists?
What is a real concern, and I really wish people understood this better, is weak expertise in AI development. One day, using AI might become very attractive and maybe even economically imperative in an increasing number of sectors. What could happen is that people with weaker and weaker AI expertise will actually be tasked with developing these AI systems as well as with their deployment and maintenance. This, coupled with the human tendency to overestimate what information technology and systems can do for us, is a risk – especially once these systems start to behave human-like.
When it comes to policymaking and AI in Europe, what picture do you see from the CLAIRE network?
It’s a constantly changing picture since many countries have recently developed digital strategies. From my work for CLAIRE and our national contact points, what’s interesting to note is that you don’t have to be big and economically powerful as a country in order to realise the enormous potential in AI, right? Luxembourg, for example, has excellent AI scientists. If the government pays attention, these countries may very well be spurred into political action rather quickly. Being a small-sized country is sometimes even advantageous because it means being able to move more quickly. Also Slovenia is excellently recognised internationally. In fact, any country can have talented people, and be a first-mover for AI.