Commission White Paper on Artificial Intelligence
Some wise person has said that if a person would travel in time from the Middle Ages to the present, we would look like gods to them with our technology. This comes to mind when reading the Commission White Paper on Artificial Intelligence, published on 12 February.
For us the promises and also challenges of AI going forward are truly impressive. But probably they are still manageable, very much like our world is manageable even though our ancestors would not believe it.
The White Paper says that ‘The Commission is committed to enabling scientific breakthrough, to preserving the EU’s technological leadership and to ensuring that new technologies are at the service of all Europeans – improving their lives while respecting their rights’. These are good ambitions in order for us to harness the benefits of this technology. Actuaries as a profession are well-placed to participate in this process.
Interestingly, the Commission has a fairly modest definition of artificial intelligence: ‘AI is a collection of technologies that combine data, algorithms and computing power’. This is very far from Alan Turing’s old definition. For Turing the real test would be a machine’s ability to exhibit intelligent behaviour equivalent to, or indistinguishable from, that of a human. While for a long time AI has promised a lot but delivered fairly little it seems reasonable for the Commission to lower the bar.
The Commission rightly realizes that the key to wide-spread positive utilization of AI is that there is trust to what is done with it. It aims at creating an ‘Ecosystem of Trust’. This requires responsible actions from those practicing AI. It needs values to avoid inappropriate differentiation among individuals. And it needs transparency to make public scrutiny possible.
The White Paper recognises that Europe is in a weaker position in consumer applications and on online platforms but stronger in many technological areas. It also holds large volumes of public and industrial data with huge potential of harnessing. If Europe can create around this an ecosystem of excellence and an ecosystem of trust, it will be a strong player in utilising AI.
The Commission aims at better flow of data across the EU. This might be interpreted that we will see similar initiatives in other sectors with those we already have for banks: PSD2 requires banks to open their interfaces to other players.
Actuaries mainly serve the insurance and pension sectors that have always been true dataholics. The tools have this far not been as good as they could have been. With technology going forward and with the implementation of the initiatives of the White Paper we will have lots of opportunities for the actuary. Will we then truly have the era of the actuary of the fifth kind in the words of Paul Embrechts: ‘data driven and model guided financial decision maker in a world governed by uncertainty’?
Most importantly however actuaries must play a central role in creating trust to the models. Without this societies and individuals will not be comfortable with artificial intelligence. We can play our role by using our central skills and obeying our code of conduct and our standards. On this basis we can create trustworthy and truly beneficial applications of new technologies. We will need to communicate what is done in such a manner that also our time traveler would be comfortable with us.
This blog is written in a personal capacity.