IGF 2021 – Day 0 – Event #125 Technology at Service of Ecosystem Restoration: How AI can Support Restoring and Protecting Natural Ecosystems Around the Globe.

The following are the outputs of the captioning taken during an IGF virtual intervention. Although it is largely accurate, in some cases it may be incomplete or inaccurate due to inaudible passages or transcription errors. It is posted as an aid, but should not be treated as an authoritative record.



>> We all live in a digital world.  We all need it to be open and safe.  We all want to trust.

>> And to be trusted.

>> We all despise control.

>> And desire freedom.

>> We are all united.

>> We all live in a digital world.  We all need it to be open and safe.  We all want to trust.

>> And to be trusted.

>> We all despise control.

>> And desire freedom.

>> We are all united.

>> Maria Andrzejewska: I'm trying to share the link with Scott who is not with us still.  Do we have Scott among us?  No?

>> Let's maybe start and in the background manage Scott.

>> Maria Andrzejewska: Okay.

>> Because we're already online.

>> It's my great pleasure and honor to welcome you at the United Nations IGF session dedicated to restoring the natural ecosystems that we have a great pleasure to ‑‑ we have a great pleasure to co-host with the Warsaw office.  I am the organizer of the event today.  This is hosted online unfortunately and we are not able to meet, but it's great to see the participants here online with us.  So I think I will pass it to Maria who is the moderator and also the director of our Warsaw office.  Maria, the floor is yours.

>> Maria Andrzejewska: Thank you very much.  I'm happy that we finally ‑‑ that we can meet today and this is my great pleasure really to open our session at this very important event, United Nations Internet Governance Forum.  I'm very honored to moderate the session.  I do hope that all our participants will join us during just a few seconds because we are still missing some of them.  Thank you very much for your presence.  This is quite unfortunate we are meeting online only.  We were supposed to be at this moment in Katowice and our experience as well as plans for future cooperations, but let's hope the future for this typo --the moment for these type of meetings will come again.

The session and presentation and dedicated to the future relation or the relation between technology and nature protection.  All in relation to supporting the UN on ecosystem restoration, which was proclaimed in the beginning of this year and this is dedicated to the improvement of the state of the ecosystem which is central to achieving the sustainable development goals, not only 14 and 15 which are about the nature but also those related to poverty health, combating the climate crisis and others.  We are looking for innovations, technological but also social, but today the advanced technologies which are in our hands are a very important part of the solution and should be integrated in plans for the upcoming year.  The goal is to discuss how cloud computing and artificial intelligence bring value to restoring and protecting natural ecosystems around the globe.

We have only one hour.  I will encourage you to ask questions in the chat or Q&A.  We don't have Q&A.  We have chat.  We'll try to answer some of the questions at the end of the meeting.  Because our time is very limited, so I'm asking Alex to start with the introduction.  Alex is a chief of big data branch and science division environment program.  He's also a chair of UN geospatial network, a number of the UN data governance group as well as a high-level committee on program strategic foresight network.  Please welcome Alex Caldas, United Nations environment program.  Alex.

>> Alexandre Caldas: Thank you so much, Maria.  I think you can hear me.

>> Maria Andrzejewska: Yes.

>> Alexandre Caldas: Thank you so much.  I'm really grateful and thank you for agreeing to Warsaw.  Thank you to Microsoft and the organizers for this important event.

Let me share you with three fundamental messages on these UN decades.  It's not about technology.  At the end of the day, it's about people ‑‑ the metaphor of the three Ps.  It's about people.  It's about places.  It's about planet.  If we look the times have changed that we are witnessing, that we are key actors today.  It's all about how much impact we can offer on people, on places, and planet.  What is the challenge here?  The challenge is come into the system that brings together people, places and planets.  And I think that's the main critical challenge of the decade for ecosystem restoration, because whenever we are addressing the impact on people, such as reduction of poverty or elimination of anger, we're also addressing the tremendous impact on biodiversity systems and land degradation and of course we are addressing the tremendous impact on climate change.  It all comes together into this very dynamic system and this systemic perspective that brings together ‑‑ in fact, all the SDGs together.

So whenever we tackle the challenge of the restoration of ecosystems, we are not only tackling nature.  We're tackling the overall system of the sustainable development goals.  And that I think is my first message.  The challenge is much more complex than we usually tend to put things together.  Because we face a system that by the end will be addressing people, places and planet.

My second message is about the key question that you asked me about is technology in the picture.  Are we using enough of the technology leverage effect to properly face direction of the challenging of people, place, and planet.  I think we're very good in identification of the technology, whether it is big data, it's cloud computing or artificial intelligence.  But when it comes together ‑‑ or black chain.  When it comes to bringing those technologies into these dependent systems, we're not so good at the effectiveness and the impact that we really want to generate.

What I'm trying to say is that we might be at the very beginning of the transformation where we provide access to the technologies and in providing access we are achieving a certain level already and we provide access.  We're still very much in the beginning of the curve of the actual use of the technology and we still have a long way to go when it comes to the actual value added that we can get from the technology to get those systems in place.

If I grasp the current status is that I will put it at the very early beginnings of accessing technology.  Still very much at the beginning.  We have some good examples of the use of technology.  We have much less examples of the actual impact of the technology on the ground.  That will be my take when it comes to the actual extensive user form of this together.

My third message is more about lasting "P" they didn't put in the picture, basically on people, places and planet.  We change with the technology.  That message is about partnerships.  Everyone is taking up partnerships as the holy grail of actually implementing things, whether it is at the ground level, local level, or whether it is at the regional level or at the global level.  The partnerships are an ingredient.  My key comment here is the full link.  We're addressing the scale of our systems through partnership, but we need to address the question of the sustainability of the partnerships.  Most of our partnerships, unfortunately, have been very much based on projects and they have a very short time frame or a short time scale and Les say two to three years whereas the problems we are tackling are long‑term problems.  We have that example.

When we try to set the partnerships with such a long time frame, we are facing a number of obstacles.  I used to summarize these challenges by what's in it for me.  So basically, if you find what's in it for me for each one of the partners that are evolving these large-scale partnerships, then you might come up with a more sustainable model.  Which are the right incentives?  What I'm trying to put as a message in the picture is even if we find the right balance to have a clear goal of people, places and planet, even if we are able to bring in the technology in a very nice way through the curve of views from access first, then the usage and then from the development of the technology, we still have a big challenge when it comes to the sustainability of the partnerships.  Because it is clear to me that most of these partnerships are related to the short come, to a project-based initiative based partnership.  Whereas everything that we've been discussing is targeting the long term.  Is targeting 10 years, 20 years, 30 years time.  So I think we should all reflect and we have the conditions and the wonderful audience in these forms to have multiple stakeholders involved to try to think the private sector, public sector, international organizations, how we can craft long‑term sustainable partnerships.  I think that's the holy grail.  Because when we go to long‑term partnerships and sustainable partnerships, then we can get the time, enough time to get these technologies on the ground and actually to go to our bottom line goal, which is to impact on people, place and planets.  I hope I have put on the table more challenge than answers sorry about that.  That comes from experience with putting these on the ground.  I hope we can have a nice and collective discussion.  Over to you, Maria.  Thank you.

>> Maria Andrzejewska: I'm muted.  Sorry.  Thank you so much for the great introduction.  The partnership and the sustainability of partnerships, that is challenging that's very important.  That's a challenging, but we have some experiences related to the network, so this is a partnership which really works for more than 30 years.  So that's possible.  And that can be successful for everyone.  But thanks for bringing those three places, planet.  Quite often when I'm talking about the climate crisis and all the challenges, I'm adding the purpose.  So the fourth "P" if, in fact, we can have it, let's also add the purpose, because that's important that we do something really for purpose.  This is really what we need.  So because our time is really very limited, so I hope we'll have time at the end to ‑‑ for some comments and questions to you directly, Alex.  But let me right now move to and invite Mr. Scott Mauvais.  Sorry for the pronunciation, from Microsoft to share with us a keynote which is titled AI for programs, how our ecosystem is supported by technology.  Scott is a director of AI and strategic partnership at Microsoft philanthropies.  Scott, the floor is yours.

>> Scott Mauvais: Thank you so much for the opening comments.  Hopefully I can address some of the challenges you put out there.  Should I share my desk top?  I've got a presentation or is someone else going to advance through the slides?  Which is better?

>> I can share.

>> Scott Mauvais: Or I can share too.  I'm happy to share it.

>> Go ahead.

>> Scott Mauvais: Do you see the presentation now?

>> Maria Andrzejewska: Yes.

>> Perfect.

>> I'm in Microsoft philanthropies.  I'm based in San Francisco.  I particularly run a program in philanthropies where we are addressing the comments on sort of the partnership side where we try and partner with some of our largest customers.  Let me advance the slide here.  Some of our largest customers around the premise with the problems we are looking to undertake are larger than any one company or S.E.C. or can take on.  You've got a list of some of the companies we've worked with.  I want to go through these quickly and give a perspective of where I sit in the organization.  And some of the nonprofits we've partnered with.  A lot of the work I do is around skilling and getting people ready for the 21st century economy, but also part of the portfolio is around AI for good.  It's 125 odd million dollar commitment from Microsoft really in five key areas.  I look after four.  AI for earth which is the one we'll talk about here.  That looks at agriculture, biodiversity, climate, water.  AI for accessibility.  Really how do we make it easier for people with disabilities to engage in everyday life, how do we improve employment and education outcomes, how do we improve quality of life.  AI for humanitarian action which is largely focused around disaster response, rights ‑‑ human rights, particularly women and children, migration, and displaced people and how do we do a lot of work around optimizing resource.  And then finally AI for health.  Also one not part of my portfolio is AI for cultural heritage.

I want to talk about the premise behind AI for earth.  We're facing extremely challenging circumstances these days.  We need to manage and model climate change.  We need to feed an increasing population while reducing the impact on our ecosystem.  The complex interplay between these systems as we try to understand them really makes it difficult to address them with traditional means.  That's where our AI for earth program comes in.  The purpose is how do we sort of merge together the world of traditional environmental science and bring together computer science to help us better engage in the ecosystem.

The premise is first how do we model these systems?  We don't have a lot of data.  It's sort of a data desert as we look at the complexity of the environment.  We have lots of data on other parts, our economic data, our healthcare systems, but water flows and water basins, biodiversity.  There isn't that rich data set.  50 how do we monitor those systems so that we can get that data?  Once we get that, we can start to build models around it and it helps us understand at human scale what it is that we need to do to better manage and protect these ecosystems.  Finally, it's that management piece.  It's that Alexander was talking about that we're not seeing a lot of impact of technology in these environments.  So how do we use AI to better prioritize our resource, help us optimize where we're going to go engage so that we can have the greatest impact.

I talked briefly in the AI for earth program about some of the areas that we're focusing on and so I wanted to drill down a little bit in each one and talk about what we mean by them.  So in agriculture, we all need to eat.  But we need to do that at a lower impact to the environment.  We need to find ways of being able to feed more and more people on less and less land.  To do that we really ‑‑ AI can help us efficiently use our resources and monitor farms in realtime so that we can increase that output.  We can do it in a less impactful way.  We can be able to do that with fewer pesticides, fewer herbicides and really help protect the environment around that.

Water.  Everything needs water.  That's the basis of life.  The demand for water over the next couple of decades is going to far outstrip the available supply.  We need to find ways to better model that water ‑‑ those water resources so we need to know where to protect them, conservative them and how do we use the water we do have more efficiently.

Biodiversity.  There is long running decays of ecosystems around us that humans are dependent upon.  They have cascading effects when we start losing biodiversity.  We can use AI to better discover and monitor and protect that biodiversity.  It's back to some of the data desert aspect and some of the examples we'll give in the next part of the conversation where we'll dive into some examples and really talk about what we're doing around species preservation.

Lastly there is climate change.  We're seeing more and more extreme weather events.  We are seeing greater impacts to the quality of life.  So what we can do with AI is help people better manage, adapt and predict these climate changes so that we can respond in a quicker way.

In terms of the AI for earth program itself there's really three pillars that we provide some grants around cash, technology, so how do we get technology in the hands of practitioners, those earth scientists, those biological scientists so they can more easily use artificial intelligence to optimize what they're doing already.  We build tools around that so we provide APIs.  We provide applications that will help people do their work more efficiently.  One example is camera traps.  We put camera traps to manage biodiversity.  It catches thousands and thousands of images.  Very few of them actually have anything in them.  A lot are blank photos of the surrounding area.  So using AI where we can strip out the 98% of images that are empty so that scientists can focus only on the particular images that have something interesting in them is a huge boost to efficiency.

Finally, in the data side.  We made available about 10‑terabytes that are available to conservation actors and we've given grants to over 700 organizations globally to help them increase their work.

In terms of making things easy to use, so the early days of using technology around environmental protection we're building what we call a planetary computer.  How do we put it easily in the hands of practitioners.  If you want to do advanced AI work, you really need to be an expert not only in your domain of science research, but you have to be fairly technically literate in understanding how to use machine learning models, how to manage clusters, how to clean your data.  We aspire to make that easier to people can focus on their domain expertise.  And so the areas that we're looking for the computer to solve is around classification, around forecasting, planning, and diagnostic.

Some examples.  We do work around land cover mapping.  How do we take satellite imagery and be able to understand what is water, what is a game trail, what is pavement, what is a building, what is that building constructed out of by looking at the roof?  In forecasting we're worked in Qatar to better understand the impacts of climate change and erosion on changing coastlines.  We'll talk more later about work we've done in the Amazon around predicting deforestation.

The planning is really where do we go do this.  We have a project with the world mosquito foundation that is trying to control fever by releasing mosquitoes that are unable to pass on fever on to humans.  What they have built a model that understands the ecosystem of each area, the economic activity of the area, the population in that area so that they can then go optimize where they release these mosquitoes.  You have to do it roughly every ‑‑ I think it's every 50 meters or so and being able to plop the neighborhoods where you'll have the greatest impact saves time.

Finally, on the diagnostic side, we are doing some work that ‑‑ how do we optimize that plan?  We're doing some work with the world's ocean foundation to attribute greenhouse gases to the shipping industry so that we can better determine what routes we can use greener fuels, what routes we can shift to battery power, which we can shift to hydrogen down the road, which ones are long haul.  We don't have good alternatives other than fossil fuel.  So how do we make sure we're optimizing the use of the fuel for the particular need and then I see Nick is on and he'll talk about around ocean mine and understanding the fishing industry.

We talked about the 700 grants.  Here is a quick map of where a lot of them have landed.  They're fairly evenly distributed.  Finally our work around carbon and climate.  I know this conference is happening in Poland, I think I would pop up a list of the grantees in Europe being some that of these might be more familiar to some of the folks attending the session.  And I think the next piece that we're going to go do is we have a project with CSI.  We've done a number of projects as part of their healthy country AI plan.  One is around protecting turtle hatchlings.  It's like 2:00 in the morning there.  Rather than participating in person but also remotely, they have prepared a video.  Is that on your side you can play?

>> Yes.  I'm going to share a screen and share the video.

>> To show important challenges in the ecosystem both on ground, management interventions.  Well.  I'm Cathy Robinson.  I work at the CSRO.  I lead and support several initiatives that provide practical examples of how to accelerate innovation with ideas of value and enable diversity.  This includes projects that focus on the ethical design and application of digital technology and AI for sustainability.

>> Hello.  I am Justin Perry from the north Australia land.  We have worked with partners to design a new software application.  These solutions enable data collection and analytical efforts to be governed by traditional owners, reflect the priority areas of concern for local indigenous communities, respond to the seasonal aspects of indigenous people stewardship of their estates and support on ground adaptive management efforts.

>> Cathy Robinson: Why?  Because indigenous people manage more than 80% of ecosystems and threatened spaces.  Over half of Australia's total land area, digital technology can help ranges manage their range.  It's critical that these tools are co‑designed to ensure that they deliver benefit pack to indigenous communities.  Our healthy country AI collaboration is helping to do just that.

>> Justin Perry: Over the past four years the program has been impairing ‑‑ to rapidly serve large areas with drone, satellite or helicopter collected footage to show important changes to the ecosystem.  This response to calls from indigenous groups are ways to design and apply technologies to solve complex environmental management problems, specifically technologies that can work with indigenous people stewardship practices and knowledge.

>> Cathy Robinson: In the national park, health country AI is being used to assess the impact of spreading of weeds on an important species on the floodplain.  With the data produced by drone monitoring, ranges have used AI to get estimates of the population.  This has enabled elders to check how affected the management has been.  The count in one wetland jumped from 50 to 1,800.

>> Justin Perry: In the northern queens land, AI has accelerated the assessment of impacts of pig control on sea turtle nesting sites.  The ranges are now able to efficiently analyze tens of thousands of images taken from helicopters to monitor predator disturbance of marine turtle nesting sites.  They have managed the pig populations leading to a 90% reduction in nest deprivation.  Healthy country AI has also been harnessed to tap into satellites and manage large herds of cattle and Buffalo using AI and space technology.  Tags have been attached to the animals and deliver realtime accurate insights into herd density, accessibility.  Healthy country AI has been used to help the Rangers predict the movement of animals which informs certain decisions about herd management.

>> The healthy country AI programs is available.  It's a great example of how indigenous led design can be ethical and useful to indigenous land managers and rapidly survey large tracts of land.  This is an end to end solution that supports an evidence‑based approach to indigenous led adaptive management of important spaces.  Our efforts to scour the impacts are now focusing on providing on ground training so that indigenous Rangers can drive and develop AI and digital tools to not only support evidence-based decisions on their country but also provide critical digital skills for future work on country.  The process will facilitate increased collaboration between the creators of AI and indigenous ‑‑ ensuring the design of AI tools and use of data supported through training courses that address gender and indigenous specific ecosystem issues, systems and priorities.

>> Justin Perry: Thanks for listening.  If you'd like to find out more, see the links or contact us on the links provided.

>> Maria Andrzejewska: Thank you for your presentation and thanks to our Australian team on what's done within the healthy country AI supporting the Rangers.  And the ‑‑ really the impressive results of the cooperation and the activities of the project presented in this historic ‑‑ it can be an inspiration for others, for other communities.  For me that's very important that it's also about supporting the indigenous people but also about increasing their digital skills, so that's really important as well.  Thank you so much and we move forward to the next presentation ‑‑ to be presented by Nick Wise, CEO of ocean line.  I don't have the title, so Nick, the floor is yours.

>> Nick Wise: Thank you very much.  I'll just get the presentation up on the screen.  Hopefully you can see that.  So I'm going to talk a bit about the ocean and the work that we do in collaboration with Microsoft and others in order to try to help protect that life in the ocean.  You may not know, but over 80% of all life on earth dwells in the ocean.  The ocean produces over half of the planet's oxygen.  At least 3 billion people rely on the ocean for their food security and one in eight people depend upon the sea to earn their livelihood.  The ocean's crucial to our level carbon cycle.  It is important for regulating our weather and our climate.  We need the ocean to survive.  But the ocean is under threat from human activity and the relentless effects of climate change.  It is stated that about 60% of the world's marine ecosystems have been degraded or used unsustainably and thousands of ocean species are threatened with extinction.  Ocean mind is a not for profit organization with a mission to power enforcement and mines to help protect the ocean's ability to pro void for human well‑being.  Our work supports regulators on the ocean priding them with training and support to increase their enforcement capacity.  We also support regulated industries to understand their compliance and take action against noncompliance.  So how does technology fit in and why is it so important?

Well, because the ocean is big.  It's really, really big.  And regulating human activity on the ocean is incredibly hard.  There's no fixed infrastructure.  No way to see.  No way to hear what's happening on the open ocean except via satellites which create a massive data set that's impossible to analyze and understand manually.  This means that AI is required to sort and filter that data, to highlight suspicious activity for investigation.  Ocean regulation is also incredibly complex with thousands of differing jurisdictions overlapping an inconsistent regulations and patchwork enforcement.  Attempting to regulate human activity on the ocean without technology is impossible.  But technology isn't just for regulators.  It's also crucial to supply chains to understand the legality and capture of seafood by using tracking data from fishing vessels, it's possible to map supply chains such as this example.  This shows all the tuna delivered into Thai ports over the course of a year by using the same underlying technology that regulators use, industry can make the approach ‑‑ can approach from opposite ends and make more rapid progress together.

Monitoring supply chains or enforcing regulated activities such as fishing is possible where this vessel tracking data is available.  But where regulations pertain to an area of the ocean such as a marine protected area, it's necessary to use remote sensing to determine vessel presence in the absence of tracking data.  And determine if that protected area has been violated.  By using satellite observations, such as synthetic radar, it's possible to detect the presence of vessels from orbit.  Ocean liners analyze over 450 million square kilometers of satellite radar in this imagery which is more than the entire surface of the ocean itself to locate vessel presence.  And that simply wouldn't be possible without machine learning algorithms to identify vessels within images.  Images can also show vessel activity.  Images from space can be low resolution or high resolution.  This image shown is a relatively low-resolution image from European space agencies sent from two satellites.  You can clearly see vessels with your own eye and see their wake paths and algorithms are being developed to detect these.  This is a high-resolution image.  Here you can see a shipment at sea in progress.  This shipment was identified using the vessel tracking data that we saw earlier and then a satellite image was taken to capture the shipment in progress.  This is a high-resolution image from digital globe at ‑‑ this was taken several years ago and it's now possible to get higher resolution images.  It's possible to see the cargo vessel next to a long line vessel.  It's possible to make identifying characteristics of the vessels such as the number of cranes.  This is really important for enforcement.

Another data source to highlight vessel activity is visible light imaging.  To illustrate this, this is a photograph that was taken from the international space station.  The green dots are fishing vessels attracting particular types of fish.  Satellites routinely gather light source information as with radar and we can correlate these detections with vessel tracking data to identify vessels or locate non‑transmitting vessels.  To illustrate every green dot is a vessel a bit like this.  Where the lights are being used to attract fish to the surface, those lights can be seen from space.  Ocean mind uses Microsoft Azure to bring together all of these data sources to understand the compliance of human activity on the ocean.  The data means that you have to use large scale computing resources, cloud resources, and AI models in order to be able to understand them, to draw out the relevant information to support enforcement.  In partnership with Microsoft, we've developed artificial intelligence models that identify specific types of fishing activity.  Those automatically identify suspected non‑compliance at scale.  The AI accurately recognizes different phases of fishing such as setting gear or soaking gear or retrieving it.  That means we can extrapolate more information from the data such as how hard is a crew being worked and whether they're having sufficient rest periods.  So all of this technology is then brought to bear to protect the ocean spaces and preserve marine biodiversity anywhere in the world.  Vessel activities are mapped and compared to all the relevant rules and regulations and they build a picture of compliance.  Their intelligence can be provided to the regulators to power up effective enforcement.

Ocean health is critical to our continues survival and unsustainable human use of the oceans is driving dramatic decline.  Destruction of whole ecosystems and the extinction of thousands of species.  But there is still hope.  Technology is an enabler for planet scale protection.  By powering existing enforcement, it's possible to turn this decline around and by the end of the decade protect the health of the entire ocean for all of our benefits.  Thank you very much.

>> Maria Andrzejewska: Well, thank you very much, Nick.  Truly impressed by your presentation.  This is like a dream.  To have access to such data and to be able to use it to support the nature protection and the ocean protection and the biodiversity and to make so many other activities.  So that's great.  The last of our presentation we have also a case about Amazon which should be leaded by the Amazon in Brazil.  She was unable to join us.

>> Scott Mauvais: I can talk a little bit about it too.

>> Maria Andrzejewska: The floor is yours.  But we have no sound.

>> Scott Mauvais: I'll wait for the sound, but we're using some satellite imagery to predict the forestation in the Amazon.  Historically society has done a good job of building dashboards and recording on it, but this organization has built a model where they look for illegal roads.  Roads that are created because they are a precursor to first some clear cutting and then burning down of the forest in order to convert to agriculture.  Should I just keep talking about it, or do you want to see if you get the audio going?  Or are the captions enough and we can read the captions?  It's possible the captions give us enough.

>> Maria Andrzejewska: You are sharing?

>> Scott Mauvais: Weronika is.  I am not.  Can you say a few words?

>> Scott Mauvais: I'll say a few words and then we can get on to the questions.  I think this runs a couple minutes.

>> Maria Andrzejewska: We have Lucia with us.  Would you like to add something?  You just joined us a second before.  You have to unmute yourself.  Unmute.  Okay.  Can you hear us?  No.  She doesn't have sound.  Okay.

>> Scott Mauvais: While she figures that out, she can correct anything when she gets done.  As I was saying earlier, historically the organizations have done a good job of reporting out the fact of deforestation and doing dashboard models which is very useful.  What we really want to help them do is take a next step and be proactive so that local authorities can intervene.  Similar to what Nick was talking about by understanding what fishing vessels are doing, and then feeding that information to authorities and then they can prioritize how to act.  How do we go about the same thing around land use management?  And as I had mentioned in the first part ‑‑ she has sound now.  I'm going to stop.

>> Hi, everyone.

>> Maria Andrzejewska: No problem.  We saw the video by Scott.  So I think that we understand what's going on and what's the ‑‑ what your project is focused on.  Maybe we move to the discussion and maybe during the discussion you can add some additional things if there is time, because I'm afraid we are running out of time.  It's 5 past 4:00 in my place.  We should finish the session quarter past, so we have less than 10 minutes for the discussion.  I have to say that a.m. really impressed by all those presentations and how the technology can support and all the activities which Microsoft started some time ago and how you are moving forward.  That's really fantastic.  So maybe I will ask you, Alex, maybe you comment, because you are looking from the UN perspective at what's going on here and what was presented.

>> Alexandre Caldas: I have to confess I'm so fascinated by the demonstrations because from Scott and from Microsoft I got this fantastic slide on how they're using artificial intelligence and to an actual demonstration on the ground.  Then the video explored.  One thing I forgot to mention, when I talk about people, place and planet, bringing in indigenous communities, this video demonstrates an actual application of doing that.  The question that comes out is how to scale these up, because at the same time we want to have maximum impact on local communities and local knowledge of indigenous communities.  How can we scale this up to that worldwide scalability model for all of these things.  Of course, it's impressive with the actual results on the ground.  Our lives depend on the ocean.  The universe beyond planet earth is another big thing because water is still there, probably, but while we do that, we have immense exploration observation of the oceans.  We have seen that with an example of Nick on the ground.  These prove that it's possible that we could see, again, on the ground how these things can be applied.  In terms of the United Nations, I think the contribution we can give to this is some kind of policy framework as well as the convening power to bring together different stakeholders.  I think these are two areas of partnerships that the UN can do, but then lead to the stakeholders around the companies like Microsoft and leave to the act ors on the ground the actual doing.  I think the United Nations can become a nice policy adviser and convening facilitator, but then when it comes to actually do it, we have to prove the actors in the discussion are a good example of demonstration of power.  Over to you.

>> Maria Andrzejewska: Thank you so much.  We are discussing today that this is so important to support the nature conservation, because the biodiversity practices and in fact that's invaluable for humanity.  But it also has a strong economy value, so ‑‑ which was also presented by the UN on restoration.  My next question goes to Scott.  What are the plans?  How would you summarize the results of what you did by the end of the 2021 and what are the next steps?  What are your plans?

>> Scott Mauvais: Great question.  I think it comes down to, you know, what he's talking about of scaling.  We're a private sector company.  We are a technology company.  What we think we can contribute to this conversation is how do we prove out the technology?  How do we make sure that it is viable?  How do we make sure it is ethical?  How can we make sure we partner with the local stakeholders around that?  Once we really look at ourselves as sort of reducing some of the riskiness of these scenarios.  And then we can pass it off to civil society who can take that and scale around that.  Success for us is where we are able to take a project, prove it out and hand it over to a private sector.  We can hand it over to the public sector so they can scale it out through their channels.  We really look through their public private partnerships and where we see ourselves sitting is in the innovation space and funding nonprofits such as Amazon, such as ocean mind, working with state agencies around building these models.

>> Maria Andrzejewska: Thanks a lot, Scott.  We have four minutes.  We were reminded time is really running out.  There was a question through the audience about are there any uses of AI technology to support the bee population and you promised to answer this question.

>> Scott Mauvais: I'm sure there are.  I'll let the other folks also chime in.  I don't know anything that's around protecting bees, but we do have a project on using bees to analyze the biodiversity in their environments.  So what we do is we essentially analyze the pollen in bees as they come in and go out of the net that understand what the flower and species are.  If there's this many of this species, there are so many insects that feed on that and birds that feed on those and we can build that biodiversity model on top of that.  It's called ‑‑ it's a French company.  They are now moving on to using that same analysis to analyze not only the species of pollen, but also the chemical makeup of that to understand what the pollutions and the heavy metals are in that environment.  So they can then build a sort of not only the biological ecosystem but also sort of contamination model for lack of a better term to help prioritize resources.  Instead of the Internet of things, they're building the Internet of bees.  I don't know if you have anything you want to chime in the last couple minutes since you didn't have a chance to sort of talk about your project.

>> No.  I just wanted to add that to the risk of deforestation, we can add many other layers and adding this to what's next and also to the bee question, we can add several other layers of biodiversity, such as carbon, such as water, so we can also understand crossing the risk of the deforestation with other layers of information, what is the risk of carbon emissions, what is the risk of loss of moisture in the atmosphere, what is the risk of loss of biodiversity.  I guess that opens a door to many other analyses that can be done in the Amazon forest and other biomes as well.

>> Maria Andrzejewska: Thank you so much.  All the time I'm looking at the clock because this is ‑‑ I'm afraid they will just close the session.  And don't give us a chance to say a word to say goodbye.  So I think maybe that's the moment when I will share with you the message.  We will prepare a summary and we'll share with you some key findings of this discussion and thank you so much for your participation.  I'm really impressed by all the presentation and I'm sure that there is a huge potential for cooperation for everyone.  Also at the global level but also here at our country level.  We'll keep in touch with everyone.  Thank you, Alex.  Thank you, Scott, for joining us.  And I think these are the last 30 seconds, so if everyone wants to ‑‑ if someone wants to say something, so that's the last moment.  I'm so afraid they will just close the session.  Thank you once again.  Let's hope we'll meet in person.