- il y a 5 semaines
Democratizing Data: Growing Food in a Planet Under Pressure
Catégorie
🤖
TechnologieTranscription
00:00Well, good morning and welcome. My name is Anna Rold. I'm the CEO of Diplomatic Courier Media Network.
00:06I'm very pleased to be here with three amazing experts in the topic that you're here to hear about, democratizing
00:15data.
00:16What does that mean exactly? Growing food in a planet under pressure.
00:21And with me today, I have Alex Orenstein. Alex is a drought specialist with GeoAgri, who provides weather data to
00:30livestock herders in West Africa.
00:33Kate Kalot, the founder and CEO of Amini, an AI startup based in Nairobi, Kenya.
00:40And Bernan Kowach is the head of the Innovation Accelerator, the United Nations World Food Program.
00:47So I wanted to give you a little bit of a context about what we're here to talk about today.
00:53And it's a bit of a paradox. And the paradox is that we live on a planet of abundance.
00:59We are able to feed the planet. That's something we've achieved.
01:03But yet millions still go hungry. So it is one of the several paradoxes of abundance and scarcity that we're
01:10dealing with.
01:11Climate shocks, inflation, fragile supply chains, they're all putting our food systems to sort of extremes, to the brink.
01:22But amid this pressure, there's also potential, especially in the untapped power of data.
01:29And so we're here to talk more about the glass half full part of the conversation.
01:36What is it out there that is happening, that's exciting, that we can harness and accelerate so that we can
01:44close this paradox?
01:47So let me begin with Alex. I'll start with you first.
01:51You work with open source tools, and it's at a time when proprietary platforms and AI, you know, the case
02:02for keeping agricultural data infrastructure and decentralized.
02:06How does that actually work in your domain?
02:13Thanks, Anna. So I think it's important to situate what we're talking about when we're talking about this agricultural data.
02:19We're talking about data that comes from satellites, weather stations.
02:23It powers the early warning systems that a lot of countries use to forecast climate shocks.
02:30And in the region where I work in West Africa, this is particularly important.
02:34And I think it's important to highlight that this push towards proprietary platforms is kind of running counter to the
02:43idea of keeping data open.
02:46From where a lot of us stand, this isn't a question of innovation.
02:50It's enclosure.
02:52You know, it's a question of an increasingly small number of corporations that are essentially trying to become gatekeepers for
02:59the data that we as a species need to survive.
03:02And this is something that's moving at light speed.
03:06And, you know, to be clear, this is also something that a lot of the representatives here at VivaTech, a
03:13lot of the institutions represented here, are playing a part in.
03:16Maybe not the beautiful people in this room.
03:18You guys are all great.
03:19But certainly throughout this conference, we have a large number of institutions present that are pushing this.
03:26Throughout the world, we've seen a retreat of public financing for open data.
03:32The number of weather stations is dropping, and the private sector is rushing to fill in the gap.
03:38Whether it's through things like Google Earth Engine or the gradual destruction of a lot of the public weather station
03:45infrastructure,
03:46it is becoming increasingly privatized and less open.
03:51So, I think it's clear to note that at this point, we're moving towards a system of climate apartheid,
03:57where countries in the global south are bearing the brunt of the climate crises.
04:03And as we continue to enclosure a lot of this data, as long as we continue to make this data
04:12less accessible,
04:13it is, we are tying their hands behind their backs in their ability to forecast climate shocks.
04:21It doesn't need to be this way, though.
04:24The case for investing in open source has been made by people much smarter than me.
04:29And I think that there is definitely a case for that.
04:33So, I'd say that the case for keeping climate data, agricultural data open, accessible, and resilient
04:40is not just strong, but, in my opinion, really the only way forward.
04:44Thank you, Alex.
04:45And you set us up in a very interesting way.
04:47Obviously, there's, you know, countries in the global south,
04:51global majority countries are feeling this pressure more so,
04:56and which was created by global, by northern countries.
05:03But the solutions are already existing.
05:06There's so much innovation already happening.
05:08So, I wanted to go to Kate next, because she is our innovation guru.
05:14I think she's actually on the ground doing the work.
05:17So, things are looking up in that sense.
05:20So, for you, Kate, the question I have is,
05:23obviously, there's a lot of hype around AI.
05:26Look, all around us, I don't think we've missed the message.
05:29What do you see as its most realistic short-term use
05:34in transforming food systems when it comes to AI?
05:39I actually think the response to this question also ties back
05:42to the first question that was asked.
05:45When we think about AI, we have to remember
05:47that it actually starts with data.
05:49You can't build any AI systems, any AI model,
05:52even if it's open source,
05:53if you don't have access to high-quality, trustworthy data.
05:58True, we can talk about open data,
06:00but what do you do when the data doesn't exist,
06:02or the data is scattered, or it's fragmented,
06:05or it's in paper format sitting in a minister's office?
06:08There is a lack of data infrastructure in the global south
06:12that is driving a lot of the data scarcity.
06:14So, even before we go to AI and open source,
06:18we have to solve for that data scarcity first.
06:20Let me give you a couple of examples.
06:22We think agriculture.
06:2470% of the world's cocoa is coming from West Africa.
06:28Think Ivory Coast, think Ghana.
06:31Yet, 85% of those farms are not geolocated.
06:34So, how are you even going to measure yields
06:36if you don't know where the farms are?
06:38And that, we call this internally the black box opportunity,
06:42where you have a lot of the food systems
06:45that still rely on the global south,
06:47but the global south is still a black box
06:49for the entire world.
06:51So, we can't actually even measure what's happening.
06:54And we, as a company, we had that naive idea
06:57that we would build an AI company for Africa,
06:59and as brilliant data scientists that we have,
07:02when we started building, we realized,
07:03where is the data?
07:05There is no data.
07:05So, that's why we decided to pivot
07:08towards building the data infrastructure
07:09for Africa and the global south.
07:12And that means really transforming our entire regions
07:16from analog to digital,
07:18looking at really structuring a lot of this
07:20scattered, fragmented, unstructured data
07:23that's in PDF, that's written, handwriting,
07:26handwritten or in Excel spreadsheet,
07:28and really doing the hard work of making sure
07:31that we can aggregate, digitize, collect all that data
07:35before we can even build modeling.
07:37So, when it comes to the AI hype,
07:40I think it's really the wrong direction
07:44for our countries to think that they will just be able
07:46to leapfrog in the AI age just like that
07:50without actually addressing the fundamental questions
07:54and the fundamental structure that have come in place
07:57to build AI systems, starting with the data,
08:00then the compute, but also underpinning all that,
08:03you have the questions around connectivity and power,
08:06because a lot of our regions still have no access
08:08to connectivity and very patchy power.
08:11Thank you, Kate.
08:12It's actually such a grounding answer from you
08:15to understand that there's this first step.
08:18Because I've been in so many meetings
08:19where experts say, well, we can just leapfrog,
08:23and that's not quite how it's going to work.
08:25I have so many follow-up questions for you,
08:27but I wanted to go to Bernard next,
08:30who is at the crux of innovation,
08:32but from the international multilateral perspective.
08:35And I wanted to ask you, I know I've talked to you before,
08:39where I've had the pleasure to know exactly what you do.
08:42So from your perspective,
08:44how do you make sure that these tech solutions
08:48aren't just well designed,
08:50but also they're designed with equity in mind,
08:54they're designed with accessibility in mind,
08:55and following up with what Kate was saying,
08:59there's barriers first
09:01before the technology actually can solve things.
09:04Yeah, it's a very good question.
09:06And I think this is where the reality of people
09:10and their lives, their livelihoods,
09:12actually really matter, right?
09:15Oftentimes when you're in a startup environment,
09:17you talk about human-centered design,
09:18lean startup thinking.
09:20Like, this is one of the principles we embody
09:22with our innovation work at the World Food Program.
09:25Most startups actually do this anyways,
09:27and if not, we would force them.
09:29But it's really about, like,
09:30how do you co-create with the people
09:33that the solution that it's intended for
09:34and not just solve a theoretical problem
09:36that nobody needs, right?
09:38So you're starting with the problem in mind.
09:40Now, as World Food Program,
09:42I mean, this is where our mission is
09:45to end global hunger,
09:47whether this is emergency response,
09:49connecting farmers to markets,
09:51across the globe.
09:53And the sad part about this,
09:55like, how do you create equity,
09:56how do you create inclusivity
09:58for people who are otherwise forgotten
10:00or left behind?
10:02I think these are the types of people
10:04that we work with, that we serve.
10:07And I can only say this
10:08where for a lot of the innovations
10:10and for a lot of the startups,
10:12when you build something
10:13that's geared towards the US
10:15or for Western Europe,
10:17you know, these solutions
10:18might look at Africa as an edge case.
10:21Think about, then, you know,
10:23global south, rural parts of the country,
10:26low-income populations.
10:28You know, you're likely not building
10:30to also include those people.
10:32And I think this is something
10:33where we want to make sure
10:35that those people in particular,
10:38they are not left behind.
10:39And I would agree with Kate there,
10:41where it's like, you know,
10:42it's like there's a big opportunity.
10:43Those people, like, if you are currently
10:46not having precision ag information,
10:48there's a lot of things,
10:50a lot of opportunity you can have
10:54to actually improve your livelihood,
10:56you increase your income.
10:57And, you know, maybe technology can help us.
11:00And I think we should talk about,
11:02like, how we get there.
11:03Well, let's talk about that.
11:05How do we get there?
11:06So we all agree that we need to ensure
11:08that there is data access.
11:10Accessibility is a big one, right?
11:12We talk about data the way back in the day
11:15we talked about refining oil, right?
11:17So we've heard all of it.
11:18The new gold, the new oil,
11:20all of that good stuff.
11:22So how do we do it?
11:24What does accessibility look like?
11:25This is a question for all three of you.
11:27Maybe we'll start back with you, Alex.
11:31Well, when we're talking about accessibility,
11:33I mean, I think it's important to...
11:37So I think we've talked a lot about, you know,
11:40Kate mentioned the idea of the lack of data
11:43that we currently have,
11:44and this is super important.
11:46You know, every single first-year statistics student
11:49learns garbage in, garbage out, right?
11:51No matter how sophisticated your model is,
11:55what you're going to get out of it,
11:56if your input data isn't good,
11:58you're going to get garbage.
12:00So I think when we look at the decline
12:03of weather stations in Africa,
12:05you know, in 1981,
12:06we had 3,000 weather stations providing data.
12:09Right now we have 300.
12:12So what have we proposed to fill that gap?
12:15Well, there's a lot of weather startups
12:17that, you know, are out there proposing new weather stations.
12:21I do believe that this is not necessarily
12:23the best path forward, though.
12:25I mean, if we start to privatize this data,
12:27and again, if we start to gatekeep this data,
12:29and require paid access,
12:32it's imposing a significant burden
12:34on some of the poorest people in the world.
12:36I mean, I'm pretty sure that most of the people here
12:38in this room take free weather data for granted.
12:41If all of a sudden you had to pay a subscription service
12:44to find out whether or not it was going to rain,
12:46well, you know, there would be riots in the streets.
12:48So I think that in terms of accessibility,
12:51we should remember that a lot of this
12:54should be treated as a public good,
12:56and that while there certainly is a role
12:57for private sector and innovation,
12:59it should not be seen as the only vehicle forward
13:02for fulfilling that gap.
13:06That's fascinating.
13:07So data is public good.
13:08So to think about sort of what are the, you know,
13:11going back to you, Bernard,
13:13you know, being in multilateralist institutions,
13:15where we've agreed as members of the UN,
13:20these are the rights we have as humans.
13:23We have right to air, right to the...
13:24Is data part of what we should consider
13:27as a public good, as a right for everybody?
13:30It is an interesting question.
13:32Like, it is, I would say,
13:34even a lot more complicated
13:35in a global or global south context in particular,
13:39where it's like, you know,
13:40having rights to data is one thing.
13:43If the data doesn't exist or isn't good,
13:44it's another one.
13:45If you do not really understand
13:47what your data privacy rights mean,
13:49that's yet another layer of complexity
13:51that you have to deal with that.
13:53Now, I think there's different...
13:55Even in this whole conversation,
13:57there's different data for different purposes.
14:00For instance, as World Food Program,
14:02we've been working a lot with governments
14:04on, like, you know,
14:04natural disaster preparedness,
14:06on, like, make, you know,
14:07flood monitoring,
14:08like, earthquake response,
14:10like, what happens after disaster strikes,
14:11because that's oftentimes when,
14:13you know, like,
14:13you have really this kind of disaster events
14:16that will all of a sudden, like,
14:18push millions of people at times,
14:20maybe into, like, poverty,
14:22into, like, homelessness,
14:23they're going hungry, right?
14:24Like, so these are the types of opportunities.
14:26Now, in terms of inclusivity
14:28for some of the other data,
14:30and I think this is where we need...
14:32I'm an advocate for,
14:34yes,
14:34if you ask me,
14:37what do we need?
14:37Do we need more money?
14:38Yes.
14:39Is there more money?
14:41Sometimes.
14:42So in absence of having more money,
14:44like, I think working with private sector,
14:45getting great startups
14:46to actually do these types of solutions,
14:48I think this is actually great,
14:50including there's things
14:51we were not able to do,
14:52like, five years ago,
14:53because AI wasn't as well developed,
14:56compute didn't exist,
14:57like, so, you know, like,
14:58so there's lots of opportunities
15:00we can do right now,
15:01and just to give you one example,
15:05like, one of the startups
15:06in our programs
15:07that's called Ignicia,
15:08what's interesting,
15:09they've been using, like,
15:11satellite imagery and AI
15:12to forecast weather data
15:14specifically for rural parts of Africa.
15:17Now, they are,
15:19essentially,
15:20there's a paid subscription model
15:21by farmers,
15:22and what's interesting there,
15:24like, if the farmers
15:25don't have a smartphone,
15:26they get a text.
15:27If they can't read and write properly,
15:28they're still a symbol.
15:29So, you're right now,
15:31and, like, it is possible,
15:32like, you are currently,
15:34maybe not fully literate,
15:35but you now have access
15:36to never being able
15:39to access before
15:40precision ag information
15:42you've never had before,
15:43and they have 700,000
15:44paying customers.
15:45So, like,
15:46it's not billions of people yet,
15:48but, like,
15:48I think there's opportunities
15:50for lots of different solutions here.
15:52And they also produce data
15:54when they're using
15:55these platforms,
15:56and so then there's
15:58a new cycle of all that.
16:00And, Kate, for you,
16:01I had the same questions.
16:03You know,
16:03what does that mean?
16:04You're on the ground
16:05actually doing this.
16:05What do you feel the need is
16:07and how to get there?
16:09I think we,
16:10I agree we need to separate.
16:11When we talk about data,
16:13different types of data
16:14will have different purpose
16:15and can be open or not.
16:17I like the idea of data
16:18as a public good,
16:19but I'll separate in two ways.
16:22Starting first,
16:23tops down.
16:24We are talking,
16:25yesterday there was
16:26the announcement
16:26about mistrial building
16:28sovereign data infrastructure,
16:30sovereign AI for Europe.
16:32So now we're coming
16:33into the questions.
16:34Governments are now
16:35waking up to the idea
16:36of data sovereignty
16:37and data residency
16:38and having to build
16:40national critical infrastructure
16:42to be able
16:42to hold their citizens' data.
16:44So that type of data,
16:46definitely not going
16:47to be open source
16:48anytime soon
16:49because there's some risks
16:50associated to it.
16:52and particularly
16:53if you look at
16:54who today is building
16:56a lot of the large models
16:58in AI
16:59and who actually
17:00owns the compute.
17:01So the risk
17:01that you would see
17:02is that a large company
17:04take your data,
17:04go and process it
17:05somewhere else
17:06and sells it back
17:07with a premium
17:08to you and your population.
17:10And in the context
17:11of Africa,
17:12today only 2%
17:14of the African data
17:15gets processed
17:16on the African continent.
17:17Let that sink
17:18for a minute.
17:19Like 88% of it
17:20actually goes elsewhere,
17:22whether it's Europe
17:23or US and such.
17:25So I think this
17:26we have to walk with
17:28and there is a caveat
17:29to this part.
17:31But then you think
17:31about bottom up
17:32and in this case
17:33I don't think
17:34we need to talk
17:35about data.
17:35I think we need
17:36to talk about information
17:37and information
17:39should be free
17:39and accessible
17:40whether it's weather data,
17:42understanding
17:43what is going to be
17:44the weather forecast
17:44over the next couple
17:45of weeks
17:46or whether it's
17:47understanding
17:48if there is a crop
17:49infestation next door
17:50or if there will be
17:51extreme rainfall
17:52in the next few months
17:54of floodings.
17:54And I think
17:55in the context
17:56of information
17:57it should be free,
17:58it should be accessible
17:58but also it should
18:00be invisible.
18:01A lot of the technology
18:03and AI narrative
18:05today has been
18:07around building
18:08large models
18:09to solve
18:09all the problems
18:10on the largest
18:12supercomputer.
18:13We actually
18:14have forgotten
18:14that AI is not an end,
18:16it's a mean to an end.
18:17It's a tool
18:18to solve
18:18a specific problem.
18:19So what is the problem
18:20we want to solve
18:22and then
18:22what is the solution
18:23that makes the most sense
18:25which data goes into it
18:26and then
18:27how can we make sure
18:28we distribute
18:28this information
18:29in a way
18:30that's accessible
18:31and digestible
18:32for the population
18:33it aims towards
18:34too, right?
18:35So farmers
18:36in the case,
18:37I think Ignatia
18:37is a great example,
18:39being able to send
18:40WhatsApp messages,
18:42SMS
18:43or even
18:43USSD prompt
18:44is amazing.
18:46We're doing something
18:46similar in Nepal
18:47where we're actually
18:49working with
18:50farming cooperatives.
18:51They're digitizing
18:52a lot of their
18:53receipts for inputs,
18:55seeds, fertilizers
18:56and such.
18:56And actually what we're doing
18:57is we're giving them
18:58information back
18:59and we're also doing
19:00a revenue share
19:01with those populations.
19:03So kind of looking
19:03into how we can
19:04bring new paradigms
19:05of circularity
19:07but also empowering
19:08more and more
19:09of this population
19:10to one,
19:12understand the value
19:14of the data
19:14but also receive
19:15the appropriate
19:16information that they need.
19:18Kate,
19:19this is very interesting
19:21to me
19:21because one of my
19:22follow-up questions
19:23is going to be
19:23how does that translate?
19:25You all alluded
19:26to co-creating
19:28with the populations
19:31on the ground.
19:32If there's issues
19:33or problems
19:35to be solved,
19:36they need to be solved
19:37by people on the ground.
19:38They understand
19:38their context
19:39much better.
19:40So just shipping
19:42a solution
19:43is not enough.
19:44So you gave me
19:45such a great example.
19:46I'd love to hear
19:47from the two of you
19:47as well.
19:48What does that look like?
19:49Bernard,
19:50maybe we can start
19:50with you
19:52especially because
19:53you're helping
19:54all of,
19:55you're incubating
19:56so many startups
19:56in that realm.
19:58So I think
19:59one of the aspects
20:00is like
20:02how do you bring
20:03the best ideas,
20:04the best minds
20:05to solve
20:05the biggest problems?
20:07And like,
20:07so any of you
20:08in the audience,
20:08if you're building
20:09a startup
20:10for a new dating app
20:11or like anything
20:12like that,
20:12that's maybe
20:13also a good problem,
20:14like keep on doing that.
20:15Now,
20:16I would advocate
20:17for,
20:18you know,
20:19like there is
20:20things on the planet
20:21where we can use
20:22the same types
20:23of like,
20:23you know,
20:23walk in the
20:25exhibition out there,
20:26like,
20:27you know,
20:27they're putting
20:28the best minds
20:29and solutions
20:30towards solving
20:31these problems.
20:32And I do believe
20:34there's an opportunity
20:35where,
20:35you know,
20:36you can solve it,
20:37you can make money,
20:38you can get investors.
20:40It is something
20:41that ultimately
20:42will make,
20:43you know,
20:43people have better lives.
20:44And I think
20:45this is like,
20:46I guess,
20:48in making a case
20:49for,
20:50yes,
20:50you should care
20:51because it's something
20:52you should be caring about,
20:53but there's also
20:54an opportunity
20:55from a business perspective,
20:56including like,
20:58when you think about like,
20:59what is the right model,
21:01there's opportunities
21:02for for-profit startups
21:03and companies,
21:04for non-profits,
21:05for the public,
21:07for collaborations,
21:08for alliances.
21:09I think this is where
21:10we should be open-minded
21:12and,
21:13I mean,
21:13I'm asking myself
21:14the same question,
21:15like,
21:15what can I do
21:16from the position,
21:18the role,
21:18the company
21:19that I currently have,
21:19my network,
21:20how can I make
21:21a positive impact today?
21:23Maybe just with
21:24one conversation,
21:26maybe just with
21:26one connection.
21:27I think if we have
21:27this mindset,
21:29you know,
21:30might be us saying,
21:31hey,
21:31we just met another person
21:33and then maybe
21:33bringing you together,
21:35that will unlock
21:36this value for people.
21:37You know,
21:38why not try that?
21:40I have follow-up
21:42questions for you,
21:42Bernard,
21:42but I'll let Alex
21:43chime in first.
21:44Thanks, Anna.
21:45I think when it comes
21:46to like,
21:47you know,
21:47making these solutions
21:48work locally,
21:50West Africa
21:50is a graveyard
21:52of agricultural
21:53information systems.
21:54over the past 20 years
21:55there have been
21:56dozens that have
21:57started and failed
21:58and the commonality
22:00that I found
22:01between a lot
22:01of the failures
22:02is that,
22:03you know,
22:04these were solutions
22:04that were designed
22:06more or less
22:07in a vacuum,
22:08right?
22:08You had your technical
22:09experts in Brussels
22:10or Paris
22:10or London
22:11who came up
22:12with these
22:12agricultural indicators
22:13that we were going
22:13to give to farmers
22:17and the ones
22:18that succeeded
22:20didn't really fall
22:22into this sort
22:22of colonial mind trap,
22:24right?
22:25This idea
22:25that we need to,
22:26you know,
22:27rather than starting
22:28from solutions
22:28that people
22:29have already developed
22:30and working with those,
22:32bringing our own in.
22:33So the most successful
22:36information system
22:37that I've worked on
22:38and I've worked
22:38on a number
22:39of failures as well,
22:40but the one
22:41that succeeded
22:41is called Garbal
22:42and it works
22:43in Mali,
22:44Burkina,
22:45and Niger.
22:47And this system
22:48provides information
22:49to farmers
22:49and herders.
22:50They call in,
22:50they say,
22:51hey,
22:51I want to bring
22:51my sheep,
22:52my cattle,
22:53to this village
22:53or this village.
22:55Hey,
22:55has it rained?
22:56And the reason
22:57why it works
22:57is because
22:57all of the indicators
22:59were developed
23:00by farmers
23:01and pastoralists,
23:03right?
23:03So, like,
23:04how that is measured,
23:05you know,
23:05whether or not
23:06you consider
23:06a strong
23:07or a weak rain,
23:08they were all measured
23:09and developed locally.
23:10We didn't come up
23:11with the indicators.
23:12We just came up
23:13with the way
23:14to use remote sensing
23:15to get the data
23:16they were already
23:17getting before.
23:19So, I think,
23:20for me,
23:21the key to success
23:22that I've observed
23:23has been to really
23:25prioritize
23:26what's already
23:27being done
23:27by the people
23:28who know
23:29the situation the best.
23:31Thank you for that.
23:32So, while you've been
23:34all talking about this,
23:36I've been thinking,
23:36okay, well,
23:37who are the stakeholders
23:38in this?
23:38Definitely the farmer,
23:40definitely multiple
23:41lateral organization,
23:42the investor,
23:44a big company,
23:45a startup.
23:46How does that ecosystem
23:47really work?
23:48How do they talk
23:49to each other?
23:51Kate,
23:52did I show you
23:52make an emotion?
23:54The short answer
23:55is no.
23:58Which is a shame
24:00because when you
24:00think about it,
24:01we all,
24:02as a world,
24:03very, very linked,
24:05right?
24:05We talk about
24:06the black box
24:07that is in
24:08emerging economies
24:09from a data standpoint,
24:10but actually
24:11what this black box
24:13causes
24:13is that a lot
24:14of the global
24:15supply chains
24:16are not able
24:17to understand
24:18and mitigate
24:18against the risks
24:19at the first mile
24:20of the supply chain,
24:21so basically
24:21where the commodities
24:22are being grown
24:23and sourced from.
24:24So when you think
24:25about the whole
24:26stakeholder
24:27from a value change
24:28standpoint,
24:29you're going to have
24:29the global companies,
24:30so the ones who are
24:31making sure
24:32that your chocolate bar
24:33is available
24:34in the supermarket
24:35so you and your kids
24:36can be happy
24:37or making sure
24:38that you can still
24:39buy phones today
24:41and that the critical
24:42minerals that are used
24:43to build your mobile
24:45or your laptops
24:47actually are available
24:49and be transformed
24:51and integrated
24:52in the right way.
24:54So these are sitting
24:55at the top of the chain
24:56and they're usually
24:57sitting in the global north.
24:58Then you come down
24:59the chain,
24:59you have the government
25:00and the government
25:01and global cells
25:02are extremely slow
25:03to change.
25:04They still run
25:05an analog system.
25:06They still don't really
25:08understand the value
25:08of the data.
25:09A lot of them
25:10are sending
25:10in the context
25:11of agriculture
25:12field agent
25:13to go and do
25:13sampling here.
25:15In Sierra Leone
25:16they come back
25:16with hard drives
25:17and then it's still
25:18very manual
25:19and still very analog
25:20which is extremely
25:22prone to error.
25:24And then at the bottom
25:24you have the population
25:26whether it's farmers,
25:27whether it's miners
25:28who still need access
25:30to information
25:30but are completely
25:31cut out of the systems.
25:32And I think
25:33a lot of the people
25:35in the world
25:36don't understand
25:37how this entire
25:38value chain
25:39relies on each other.
25:41Every single actor
25:42across the board
25:44has a role to play.
25:45And around that
25:46you have innovative startups,
25:48you have large
25:49cloud service providers
25:50that now will launch
25:52a small language model
25:53in Swahili
25:55or any other
25:56African language
25:57not always taking
25:59into account
26:00what is happening
26:01on the ground.
26:01And you have a lot
26:02of European startup
26:03or American startup
26:04truthfully who say
26:05we're going to solve
26:06Africa's problems
26:07and we're going to
26:08build technology
26:09that will work
26:10everywhere.
26:10They're telling you
26:11that they have a model
26:12that's 90% accurate
26:13across the entire world
26:15and I can tell you
26:16they're actually lying.
26:17So we have to have
26:18a big reality check
26:19and I think
26:20the entire ecosystem
26:21has to work together
26:22because, you know,
26:24these risks
26:25are not just a risk
26:26for local population.
26:27They're a risk for you,
26:29they're a risk for your family
26:29because a lot
26:30of the food systems
26:32still rely on population
26:33in the global south
26:35to be able to deliver
26:36what's on your plate today.
26:38So we have to make
26:38an effort
26:39and this entire ecosystem
26:40has to work together.
26:42Very powerful.
26:45Yes,
26:46that's an applause moment.
26:47Thank you.
26:51Bernard,
26:51why didn't you guys
26:52fix this at the UN?
26:54So I think
26:55this is where
26:57so the sad part
26:58is like
26:58as you said
26:59in your introduction
27:00there's enough food
27:00on the planet
27:01to feed everybody
27:02to live a healthy life.
27:05Now,
27:06there is different
27:06root causes, right?
27:07Like so right now
27:09343 million people
27:10marching towards salvation
27:11which means
27:12root cause there
27:13is like conflict,
27:14wars,
27:14you know,
27:15climate-related disasters,
27:18it's like inflation,
27:19it's poverty.
27:20Like there's reasons
27:21why, you know,
27:22it's a solvable problem
27:23but it hasn't been solved.
27:25Now,
27:25you could fix it
27:26with money
27:26which,
27:28you know,
27:29and I encourage
27:29governments,
27:31private sector,
27:33individuals,
27:33like we have an app
27:34called Share the Meal,
27:35you can donate 70 cents
27:36to share your meal.
27:38Like it's easy things
27:39you can do.
27:39Like you can contribute.
27:41No,
27:41money could fix it
27:42but it's not.
27:43So if money
27:44is not the solution
27:45or it's not coming
27:46then we need
27:47to think differently.
27:47So like
27:48we can't just be
27:49hopeful thinking here.
27:51now,
27:53we know
27:54that there's
27:55different systemic issues
27:57that we're talking about.
27:58Like data,
27:59like different aspects
28:00is one of them.
28:02Transportation,
28:03supply chain,
28:03transparency,
28:05like rule of law,
28:06there's different things
28:07that can actually be changed
28:08that are sometimes
28:09at this cusp of that
28:10and I would also argue
28:12there's an opportunity
28:13for,
28:14you know,
28:15new ways of thinking
28:17or rethinking
28:18these aspects
28:18where it's like
28:20instead of saying,
28:21well,
28:22you know,
28:22I am at the company,
28:23I need to sell my products,
28:26you know,
28:26maybe there is an opportunity
28:27to sell your products
28:28also to currently
28:29low income communities
28:31that are not
28:32the highest profit margin
28:33for you
28:33but,
28:34you know,
28:34maybe they're
28:35your future customers
28:36and I think
28:37there's an opportunity
28:38there to really
28:39just like
28:40take off
28:41this kind of
28:42dogmatic thinking
28:43of like,
28:44you know,
28:45this group is bad,
28:46this one is good,
28:47we do need to collaborate better,
28:48we need to come together more
28:50and then also
28:51look for the individuals
28:53that are actually
28:54willing to,
28:55you know,
28:56champion change.
28:57I think so many people
28:58want to do something
28:59and they don't know how
29:01and I think
29:02that's an opportunity
29:03right there,
29:03it's like,
29:04you know,
29:04maybe in your company,
29:05your company doesn't have
29:06like the biggest effort
29:08to change something
29:09but maybe you
29:10wanted to contribute,
29:11maybe you
29:11and somebody else
29:12and I think
29:14there's an opportunity
29:15right there
29:15to foster
29:17that type of exchange
29:18including like,
29:19you know,
29:20having a conversation
29:21about this,
29:23AI is like
29:25one of the most
29:28underappreciated aspects
29:29that will change
29:30the realities
29:30of like billions
29:32of people
29:32and we're not
29:34talking about this
29:34and like this is
29:35where it's important
29:35that we also bring
29:36this mainstream
29:37that's like,
29:37okay,
29:38we need to talk
29:38about the people,
29:40we need to think
29:41about like the livelihoods,
29:42the lives
29:42that they can lead
29:43in a better future
29:45rather than one
29:46that's dystopian.
29:47Bernard,
29:48thank you
29:48and I didn't mean
29:49to put the onus
29:51on that you went
29:51to fix everything,
29:52however,
29:52I did want to mention
29:55institutions tend
29:56to be very resilient
29:58to change
29:59because we built
30:00them that way,
30:01however,
30:02we humans
30:04are not that way
30:04at all,
30:05we're actually
30:05the reason
30:06we've survived
30:06thousands of years,
30:07not to go back
30:08to Darwinian laws,
30:11but we survived
30:12because we're agile,
30:13we know how to change
30:14in crisis moments
30:16and that,
30:16so putting
30:17the individual agency,
30:20allowing the individual
30:21to reclaim that power
30:23to shift institutions,
30:24that's actually
30:26a very powerful message
30:26and I think
30:27that's something
30:29World Food Programme
30:30has done
30:31with your innovation arm,
30:34so that's to say,
30:35I didn't mean
30:36to put you
30:36on the spot there.
30:38No offense,
30:39look,
30:40it's like where
30:40I think one of the aspects
30:42is like,
30:42it actually is a
30:43solvable problem,
30:44like this is what
30:44is said,
30:45right?
30:46And at the same time,
30:47if it's not like,
30:49you know,
30:49just asking for money
30:50and say like,
30:50give me money,
30:51like please do,
30:53but then in addition,
30:55let's also rethink
30:56technology,
30:56data,
30:57like everything
30:57we talked about.
30:58So yeah,
30:59so don't stop
31:00the money influence,
31:01I think that's
31:02very important
31:03and I think
31:04we're talking about
31:05just like with data
31:06and precision,
31:07we have to be smart
31:08about it,
31:09like so if the solution
31:10is not just money,
31:11what else is out there
31:12that we need to think about?
31:13So I wanted to go
31:14sort of for our final round
31:16to ask a question
31:17of all three of you,
31:19is not the,
31:20when we talk about
31:21the future,
31:22we're thinking like
31:23way far ahead
31:23and when we talk
31:25about AI,
31:25the future is in
31:2618 months,
31:29so do you feel
31:30that this discussion
31:33we've had today
31:34around data scarcity,
31:35equity,
31:36accessibility
31:36is one that is
31:38a glass half full,
31:40a hopeful discussion,
31:41what gives you hope
31:42about this very
31:44near term future?
31:45And we'll start
31:46with you Alex,
31:47we'll go back
31:47to you Kate
31:48and then Bernard.
31:50Thanks Anna,
31:50I mean I think
31:53it's a double edged
31:54question right
31:55because on one hand
31:55from my perspective
31:57the future is,
31:59the immediate future
32:00is bleak
32:00when it comes
32:01to how data
32:03is going to be
32:04able to be accessed
32:05for people
32:05in the global south.
32:07However,
32:08one thing that does
32:08give me hope
32:09is that there is
32:10an increasing
32:10understanding
32:11that this crisis
32:12is fundamentally
32:13a political one
32:14and that it is
32:15one of choices
32:16and the old narrative
32:18right,
32:19that climate change
32:20was just something
32:21that happens
32:21to us in a vacuum
32:23is gone
32:24and so I think
32:25that the movement
32:28that we've seen
32:28all around the world
32:29of people
32:30gradually understanding
32:31that it is a political
32:32move,
32:32it is a political question
32:33is giving me
32:34a lot of hope
32:35for our ability
32:36to reframe the problem
32:38and address it.
32:39Thank you Alex.
32:41KG, we're next.
32:42I think
32:42on my side
32:44we kind of need
32:45to deconstruct
32:47and reconstruct
32:47the way we look
32:48at innovation
32:49and what it actually means.
32:51There's a lot
32:52of biases
32:52that have been built
32:53from global north
32:55the US
32:56around AI
32:57that we now
32:58need to
32:59really unpack
33:00and go back
33:01to fundamentals
33:02and what gives me
33:03hope is that
33:04we're now seeing
33:06a lot of innovators
33:07coming back
33:07to their regions
33:08or their regions
33:10of origin
33:10whether it's Africa
33:11Latin America
33:12Southeast Asia
33:13and not just building
33:15take a naval solution
33:16not just building
33:17I'm going to build
33:18one model
33:18that will do
33:19one thing
33:19they're actually
33:20building critical
33:21infrastructure
33:22critical infrastructure
33:23onto which
33:24others will be able
33:25to build
33:25and I think
33:26for me
33:27this is where
33:28I see the biggest hope
33:29and the biggest change
33:30in the next coming years
33:32we're going to start
33:32seeing more and more
33:33bottom up
33:34local
33:35localized
33:36but also
33:37critical solutions
33:38native
33:39on local infrastructure
33:41Thank you Kain
33:42Bernard
33:43What gives me hope
33:45is that
33:46I see a lot of
33:48like opportunities
33:49solutions
33:50startups
33:51innovations
33:51that are
33:52right now
33:53being built
33:54that were impossible
33:56a couple of years ago
33:58and this is something
33:59that we should
34:00take to heart
34:02and think about
34:02like what more
34:03can be done
34:04right
34:04like so there's
34:05all of a sudden
34:06opportunities
34:07for people
34:07who have been
34:08left behind
34:09for maybe
34:10sometimes generations
34:11to really now
34:12be able to be part
34:13of a global community
34:14be part of a global
34:16digital ecosystem
34:17now have access
34:18to data
34:18that it was too expensive
34:20to do this
34:21like on a manual base
34:22but now all of a sudden
34:23AI enables this
34:24kind of leapfrogging
34:25it's not
34:26you know
34:26the silver bullet
34:27the one solution
34:28for everything
34:29but I do have hope
34:31that's possible
34:32and so many
34:33like companies
34:34AI engineers
34:34say hey
34:35how can I help
34:35and I think
34:36that gives me hope
34:37thank you all so much
34:39I think the
34:39one big takeaway
34:41for me
34:41is
34:43we can change
34:44if we change
34:46the systems
34:47and systems
34:48are made
34:48by people
34:49and people
34:50can change
34:51easily
34:51or we think
34:54that they can
34:55or that
34:55we want to change
34:56but I really
34:57appreciate
34:58the work
34:58that you're doing
34:59to advance
34:59a more sustainable
35:01food security
35:02system
35:03and thank you all
35:04for your attention
35:06applause
Commentaires