 Artificial intelligence or machine learning have made huge advantages over the last years. We have lots of data available, the models are getting better, the science and also the compute power is there. In the area of health, for example, we've seen big advantages in the area of, for instance, diagnostics. When you make an AI experiment in the health area, you have a certain source of data and you are setting it up and so you come up with the results. But what you actually need is a community approach where everybody gets together and everybody contributes in an ethical way data to the machine learning experiment and then it's conducted and then you publish the result and make it available to everybody else. So the strategy is to exactly set up this community approach. We try to bring together clinical doctors or public health experts that provide us with data from the field with problems that need to be solved that are relevant. We try to create this data to set up the machine learning AI problem. Once we have that, we will call for solutions for AI solutions to this problem and publish the whole evaluation and that hopefully gets then taken up by the regulatory bodies around the world or by public health institutions. So at the previous meeting, we had two so far. At the first meeting, we basically set up the whole structure. At the second meeting, we basically got responses to our call for use cases and we did adopt eight use cases that we are further considering. Up until this meeting, we have actually received more information on these use cases and these use cases could maybe now enter a second stage, some of them at least, where we do a draft benchmarking and do a dry run of our process that we have developed so far. The focus group in 2019 is to create a number of benchmarks that show that our approach to the evaluation and to the setup of community-based AI for health methods is actually working.