 We now turn to Anton Ellenbrook of FAO to reflect on the forum's outcomes in relation to FAO's Fisheries and Agriculture Information and Knowledge Management Agenda, and the need for future collaboration processes for FAO to support. The Information and Knowledge Management Team, FISHINFO, provides access to data and IT services, promotes digital innovation, and connects stakeholders. So I'll hand over to Anton if we could please, video please. Hello everybody, my name is Anton Ellenbrook. I work in the Fish Info Team of the FAO Fisheries and Agriculture Division. And today I will quickly introduce you to what we have done in FISHINFO team with AI, but maybe more important, I will explain what happens at the higher level in FAO with the Rome call. The objective is to leave no one behind and if we talk about the 35 million fishers on the planet, plus the people that work in agriculture, plus the people that work in processing and trade, we talk about a really significant number of people that stand to benefit but also have some risk when we talk about artificial intelligence and fisheries and agriculture. So the FISHINFO team, we manage data and IT services for the NFI division, but we also work with the direct involvement in digital innovation. We are active in projects, we are active in collaboration with other teams towards a digital future. So another aim is to connect stakeholders in the blue transformation. So we all know that the digital transformation is upon us. So how do we make sure that this also benefits people that have a lot to gain from improved technologies and fisheries and agriculture. Our team also holds a secretary of several partnerships, for instance in fisheries and agriculture domain, but also people that work in knowledge management and document repository, for instance, YASFETI. We're also active in statistics and geo info data harmonization and standardization, and that is then visible in various applications, like the firms and fisheries and resources monitoring system, of which you see an example in the bottom right of the screen, where you should see all the harmonized and standardized fisheries data accessible through maps, but also accessible through repositories and other data entry systems. So a few examples of what we do with the FISHINFO team in terms of digital innovation. So on the right you see already an example of how we detect fisheries and agriculture. This was an experiment done with Google TensorFlow. We also work to identify coastal land cover and classify what we see in the images for instance using a random forest approach. On the bottom right you see an example of a land use classification over south Sulawesi, and the confusion matrix that comes with the We also try to monitor some human activities with remote sensing, but there it is obvious that there are lots of questions surrounding artificial intelligence. So what can we see for instance in agriculture case occupancy, we can really detect and see what people are doing in a individual case. But they said the information that they really want to get from their field and use an FFO or other organizations to monitor agriculture. Don't we have a lot of artificial intelligence ethical problems there. Other options that we are currently discussing is options in fisheries. So what can we see from remote sensing in fisheries but also there are a lot of questions surrounding the proper use of technologies, including artificial intelligence in monitoring of what's happening on the open oceans and these themselves. So we do not try to answer these questions by ourselves so the team has a lot of collaboration so a few of the examples that we have in the collaborations ongoing at the moment are in a blue cloud project which is a horizon 2020 project where we only work in fisheries and agriculture, but there are other teams, not an FFO network for instance on plankton image classification models, but also plankton distribution models that start with the genome sampling in the Atlantic Ocean, but and people that work on essential ocean variables. In all three cases there is a lot of artificial intelligence to increase the knowledge that we have on the oceanic systems by applying AI on samples and location based observations. We also collaborate with NFI teams and I think this seminar is a fantastic example of a cross team collaboration here more focused on image analysis, then some examples of images on the bottom right. So the quick team introduction so what happens at an FFO level, there's one important initiative going on that is the Rome call on artificial intelligence, leaving no one behind. The Rome call for artificial intelligence was assigned on the 28 of February of 2020 and it's a joint declaration of the food and agriculture organization of the United Nations FEO, the Pontifical Academy for Life IBM and Microsoft with the United Nations Open to others to join the same initiative. It's a high level agreement on the use of AI in the food value chain, and underscores the ethical importance of properly contextualizing AI activities to leave no one behind. So what is actually contained in the Rome call it has six main principles that I will quickly read out so this about transparency and principles AI systems must be explainable to users and the wider public. We need to be inclusive so we need to take care of all the human beings and make sure that everyone can benefit and all individuals can be offered the best possible conditions to express themselves and develop themselves. We need to have responsibility in the system so those who designed and deploy the use of AI systems was proceed with responsibility and transparency. We need to be impartial, we do not want to create or act according to bias, the safeguarding fairness and human dignity and all assistance that we deliver. There must be reliable so all AI systems must be able to work reliable, not only today but we have already seen some presentations are how do we transfer from one situation to another that is a good example of that we need reliable AI systems. We need to provide an offer security and privacy so AI systems must work securely and respect the privacy of users I already explained to you how did we really meet this limitation already if we start to monitor fisheries and agriculture with remote sensing technologies. That's a really valid inclusion in the school. So for me, I believe this conference showed a lot of awareness around the topic of the call. So let's seize the opportunity to hear this call of the wrong goal and build a community that makes AI work for the benefit of all. Thank you very much for and hope that a good meeting. Thank you very much Anton and thanks for sharing FAO fisheries community and global partnership work and its development of acceptable norms. I think this is really one of the key messages coming out of the meeting. We've, we've heard about some of the key technical developments that we're working on whether it's trying to identify species whether it's trying to measure species and the number of presentations in that regard will be made available later for people who are on time zone. We've also heard about some of the social issues obviously we cannot just develop algorithms and expect the answers to flow and even if the answers do flow whether they'll be acceptable. And we need also heard from most presenters that it was a case that they needed all pillars to keep up the overall ecosystem of AI, a bit like what Hasan was mentioning earlier and how we bring all these pieces together. And I think as with the number of presenters coming from small teams based all around the world how do we share the wins and also share the failures so that we can help each other reach extend our reach and allow us to develop at a speed which is required in under the requirements that we have on us today.