AI in Africa: Beyond the algorithm

9 Jul 2025 10:40h - 11:00h

Session at a glance

Summary

Kate Kallot, founder and CEO of Amini, delivered a keynote address about the urgent need for AI equity and sovereignty in the Global South. Drawing from her personal family history in the Central African Republic, where her grandfather was killed by a dictator for refusing corruption and her grandmother continued educating her children in prison, Kallot framed AI development as a matter of justice and representation. She highlighted critical disparities in AI infrastructure, noting that while the US and China hold 90% of global data center capacity despite representing only 21.6% of the world’s population, Africa has less than 1% despite comprising 19% of global population. Kallot argued that current AI systems are fundamentally exclusionary, as they train on internet data while 2.6 billion people remain unconnected, making entire populations invisible to these technologies.


She emphasized that talented developers across the Global South are being relegated to data labeling roles rather than being empowered to build AI solutions for their communities. Kallot presented examples of local innovation, including M-Pesa’s mobile banking success in Kenya and Clinton, a Nairobi data scientist now leading efforts at Amini. Her company is building data infrastructure for the Global South based on principles of community engagement, open source development, and efficiency through smaller, localized AI models. Kallot advocated for a collaborative south-to-south framework that would enable countries to resist inappropriate AI solutions while embracing innovation that serves their specific needs. She concluded by asserting that the true AI revolution is emerging from cities like Manila, Nairobi, and Jakarta rather than Silicon Valley, representing a moment for the Global South to author its own technological future.


Keypoints

**Major Discussion Points:**


– **AI inequality and the global digital divide**: The speaker highlights how AI development is concentrated in the US and China (90% of data center capacity), while regions like Africa (19% of world population) have less than 1% of compute infrastructure, leaving billions disconnected and invisible to AI systems.


– **Historical context and personal motivation**: Kate Kallot shares her grandfather Joseph’s story – a brilliant African police officer in the 1950s who was killed by a dictator for refusing corruption – to illustrate how talent and resources in the Global South have been systematically suppressed, connecting past colonial extraction to current AI exclusion.


– **Moving from extraction to regeneration**: The discussion emphasizes shifting from using the Global South merely as a source of data and labor (like data labelers) to empowering local communities to build their own AI solutions that serve their specific contexts and needs.


– **Amini’s approach to AI sovereignty**: The speaker outlines her company’s strategy of building data infrastructure for the Global South through local-first principles including community engagement (supporting farmers in Nepal), education partnerships, and efficient local models rather than chasing the largest global systems.


– **Collaborative framework for AI revolution**: The call for South-to-South collaboration, rejecting the “catch up” mentality, and positioning the Global South not as followers but as leaders of an inclusive AI revolution originating from cities like Manila, Nairobi, and Accra.


**Overall Purpose:**


The discussion aims to challenge the current AI development paradigm that excludes the Global South and present a vision for AI sovereignty where developing nations author their own AI futures rather than being passive consumers of Western technology.


**Overall Tone:**


The tone begins with urgency and concern about systemic exclusion, transitions through personal and historical narrative to build emotional connection, then shifts to empowerment and hope. The speaker moves from highlighting problems to presenting solutions, ending on an inspirational and revolutionary note that positions the Global South as the true center of meaningful AI innovation.


Speakers

– **Kate Kallot**: Founder and CEO of Amini, a startup focused on building data backbone for the global South. Previously worked for big tech companies including Intel, Arm, and NVIDIA for more than 15 years.


– **Moderator**: Role involves introducing speakers and facilitating the discussion. No specific expertise or title mentioned beyond moderating duties.


**Additional speakers:**


None identified in the transcript.


Full session report

# Comprehensive Report: AI Equity and Sovereignty in the Global South


## Executive Summary


Kate Kallot, founder and CEO of Amini, delivered a keynote address challenging the current paradigm of artificial intelligence development and its exclusion of the Global South. Drawing from personal family history and technical expertise gained from working at major technology companies including Intel, Arm, and NVIDIA, Kallot presented an argument for AI sovereignty that moves beyond traditional narratives of technological catch-up to embrace indigenous innovation and community empowerment.


Kallot’s central question framed the entire presentation: “What does it take for the global South to not just be connected to an AI future, to not just be carried into an AI future, but actually authoring its own AI future?”


## Key Arguments and Themes


### The Systematic Exclusion of the Global South


Kallot’s central thesis was that “we are living through a time where entire regions are at risk of being left out of the future. And that’s not happening by accident. It’s actually happening by design.” She presented statistical evidence highlighting three critical divides:


**The Connectivity Divide**: With nearly half the world’s population lacking internet access, AI systems trained on internet data fundamentally exclude 2.6 billion people. “While those models are using data that’s on the Internet to train the systems, and you have half of the world’s population that is not on the Internet, who does it actually represent? It doesn’t represent our systems, our knowledge systems, our culture, our languages. It doesn’t represent us. We invisible to those systems.”


**The Compute Infrastructure Divide**: 90% of global data centre capacity is held by the United States and China despite representing only 21.6% of the world’s population, whilst Africa maintains less than 1% of capacity despite comprising 19% of global population.


**The Representation Divide**: Current AI systems fail to capture the knowledge systems, cultural practices, and linguistic diversity of the Global South.


### Historical Context and Personal Motivation


Kallot shared her family history to illustrate the long-standing suppression of talent in the Global South. She recounted the story of her grandfather Joseph, a police officer in the Central African Republic during the 1950s who was killed in 1969 by dictator Bokassa for refusing corruption. Her grandmother Christine was left under surveillance residence at the Sudan border with five children.


This historical lens framed contemporary AI development as part of a continuum of extraction, with Kallot noting that “today we’re standing on the edge. We’re standing on the edge where AI can be the final pin to erasing centuries of our knowledge systems, of our culture, of our languages. Can deepen the extraction that we have seen in our countries. Or can be the most inclusive leap in history.”


### Innovation Through Constraints and Local Solutions


Kallot argued that “constraints has always been our superpower. Specifically, constraint-driven creativity. It has always forced us to think differently. It has always forced us to go against the tides.” She illustrated this with the story of William, “the boy who harnessed the wind,” who built a windmill in Malawi during a drought.


She also cited M-Pesa, Kenya’s mobile banking system created in 2007, emphasizing that “technology, to truly be useful, it has to be invisible” and should integrate seamlessly into existing contexts.


Kallot highlighted Clinton, a data scientist from Nairobi who is now “head of data science at Amini and is actually sitting right here with you in the audience,” as an example of local talent empowerment.


### Building AI Infrastructure for the Global South


Kallot outlined Amini’s approach through community-centered development and local partnerships:


**Community Engagement**: Working directly with farmers in Nepal on spices like turmeric and ginger to digitize and build representative datasets while ensuring value remains within communities.


**Educational Partnerships**: Collaborating with local universities and partnering with “Zindia Africa” to train developers on geospatial data handling and AI tools.


**Local Infrastructure**: Supporting the Barbados government to create a local R&D hub with microdata centres powered by renewable energy.


### Strategy for AI Sovereignty


Kallot proposed a framework built around local-first principles:


**Resistance**: “We resist AI that doesn’t serve us.”


**Innovation**: “We innovate AI that serves our communities.”


**Selective Adoption**: “We embrace AI that can be adapted to serve us.”


She emphasized that “the question was never whether we would catch up. It was always whether we will be able to use AI as a true equaliser.”


### Vision for the Future


Kallot concluded with a call for South-to-South collaboration and a vision of the “true AI revolution” emerging from “Manila, from Bridgetown, from Nairobi, from Accra, from Jakarta, and from Bangui” rather than Silicon Valley. She positioned this as a defining moment for the Global South to “author its own technological future.”


## Key Insights


Kallot’s presentation reframed several conventional assumptions about AI development:


– AI exclusion happens “by design” rather than by accident, challenging narratives of inevitable technological progress


– Current AI systems trained only on connected populations are inherently incomplete and biased


– Constraints can serve as “superpowers” that drive superior innovation rather than merely acceptable alternatives


– The Global South can lead rather than follow in creating more inclusive AI systems


The keynote provided both a diagnosis of current AI inequality and a practical roadmap for addressing these challenges through community-centered development, appropriate technology solutions, and collaborative frameworks that prioritize sovereignty and local benefit over external extraction.


Session transcript

Moderator: founder and CEO at Amini. It’s absolutely fabulous to have her on the stage. I’ve seen her speak before. Please join me in welcoming a pre-lunch welcome, please, with slightly more coffee. It’s the fabulous Kate Kallot.


Kate Kallot: We are living through a time where entire regions are at risk of being left out of the future. And that’s not happening by accident. It’s actually happening by design. Where AI has become the default interface for millions, billions are still at risk of being left out. Good morning, everyone. My name is Kate Kallot, and I’m the founder and CEO of Amini, a startup that I founded about three years ago after spending more than 15 years working for big tech, starting at Intel, then Arm, then NVIDIA. And when I founded Amini, it was not to chase the latest technology trends. It was not to build artificial general intelligence or whatever that means. It was actually to answer one simple question. What does it take for the global South to not just be connected to an AI future, to not just be carried into an AI future, but actually authoring its own AI future? And today, for this talk, I would like to take you down memory lane. Going to Central African Republic, Bangui, back in the 1950s, my grandfather, Joseph, was born under a French colonial regime. He was one of the first African graduates of Saint-Cyr, the national police school in France. Brilliant student, he went on to studying at Indiana Police Academy, at Washington Police Academy, and even met Kennedy. He went on to becoming an Interpol officer traveling the world. Think about it. African in the 50s, traveling the whole world, having a successful career, yet there was one question that was bothering him. How could our country, who is extremely rich in natural resources, extremely rich in talent, doesn’t actually develop itself like the places that he has visited? So he decided to do one thing. He decided to come back to CAR and build and change the country from within. He went on to becoming chief of national police in the country. And in 1969, he was killed by the dictator of the day, named Bokassa, who couldn’t fathom and was fearing that the people who had not given to the corruption that he was setting up would take him out of power. After Joseph was killed, my grandmother, Christine, was put in surveillance residence, aka prison, at the border of Sudan with her five kids. But while there, while they were cut out of school, contacts with family, contacts with friends, she decided to hide learning textbooks into cassava bags so she could actually continue giving her kids an education and continuing to educate them over time. My grandparents taught me the meaning of justice for the first time. And their defense is the reason why I’m standing here today. And I have a question to ask you. Whose freedom is today’s AI actually serving? Whose it is serving where we still have a global connectivity divide, while close to half of the world’s population still do not have access to Internet? Whose it is serving when we still have a compute divide? When you think about the infrastructure that’s needed to train AI systems, data center, 90% of the world’s data center capacity still sits in two countries, US and China. The rest of the 10% mostly concentrated in Europe. While those two countries only contribute to 21.6% of the population. Take in contrast Africa, 19% of the world’s population, less than 1% of the world’s data center capacity. And it’s the same across the global south, in the Caribbean, in Latin America, and in Southeast Asia. Whose it is serving where while you have 500 million of users on the most popular AI chat bot today, you still have 2.6 billion people that are unconnected. And unconnected for me truly means invisible. While those models are using data that’s on the Internet to train the systems, and you have half of the world’s population that is not on the Internet, who does it actually represent? It doesn’t represent our systems, our knowledge systems, our culture, our languages. It doesn’t represent us. We invisible to those systems. Whose does it serve when our use from Kenya to the Philippines, who are digitally native, brilliant developers, have studied computer science, are still being left at the bottom of the AI value chain. While they’re being constrained as data labelers. And not given a chance to capitalize on an opportunity to apply their skills to build things that matter for their regions. Today we’re standing on the edge. We’re standing on the edge where AI can be the final pin to erasing centuries of our knowledge systems, of our culture, of our languages. Can deepen the extraction that we have seen in our countries. Or can be the most inclusive leap in history. We are at a place where AI can disrupt power structure, but in reality has been entrenching the status quo. So what do we need to do? We need to move from adversity to ingenuity. We need to move from scarcity into innovation. And we’re not going to win by chasing the largest model or the largest supercomputer. But we will win by harnessing our digitally native use and also supporting our communities. And we are not stranger to doing so. There’s been some very heartwarming examples. Like William, the boy who harnessed the wind, who built a windmill to power his family home in Malawi during a drought. Or even more well-known ones, M-Pesa, who was created in 2007, long before the WeChat pay, the Apple pay of this world. And M-Pesa was created to solve one specific problem. The lack of banking infrastructure in Kenya. And today it’s being used by more than 96% of the population. M-Pesa taught us something. Technology, to truly be useful, it has to be invisible. We also have some examples that are less known. Like Clinton. Clinton is a brilliant data scientist from Nairobi. And long before he actually graduated, he had built already three startups to solve really from the University of Michigan. He is a professor of computer science and he is a leader in solving problems that are around him for his communities. Think about deforestation, think about monitoring or wildlife conservation. He went on to being a leading figure in the tiny machine learning ecosystem, both in Africa and worldwide. And Clinton today is the head of data science at Amini and is actually sitting right here with you in the audience. So, how do we create more Clintons? How do we empower them to lead? How do we give them the tools to build the future of their communities and their countries? Constraints has always been our superpower. Specifically, constraint-driven creativity. It has always forced us to think differently. It has always forced us to go against the tides. So, how do we turn the tides today? You know, before, building a critical infrastructure meant building a road, building a hospital, building a bridge. Today, it actually means building the technology infrastructure for the Clintons of your country to be able to prosper. Connectivity is not just about access. It’s about agency. Compute is not just about having GPUs in your backyard. It’s actually about enabling your use to produce intelligence. Localized intelligence. Data should not just be a mean to train a large language model. It should actually benefit the people who are generating it and should serve them. And that’s what we’re building at Amini. We’re building the data backbone for the global south. Not just as an extractive layer, but actually as a regenerative one. From the connectivity, the infrastructure, and the computes. Making sure that our countries are not just participating in AI, but leading their own AI revolution. We’ve turned systemic fragility into national-level infrastructure. And we’re bridging the gaps of frontier technology in the global south. Addressing the questions of data sovereignty and residency. Of hyperlocalized intelligence. Sovereign compute infrastructure. But also climate resilience. From working with local universities, governments, and communities, and engaging them at every single layer of the ecosystem. We were born in Africa, but we’re truly serving the global south. And we’re doing this with local-first principles in mind. The principles that have shaped our cultures for the past centuries. Starting with community. Enabling community. Not just technology community, but the broader community. Starting with farmers in Nepal, we’re supporting farmers to digitize themselves. Enabling them to build data sets that are truly representative of their local context. Farmers and spices, turmeric, ginger, and really supporting them to build their own data sets. And we’re also supporting them to build their own data sets that are truly representative of their local context. And we’re also supporting them to feed into the systems while keeping the value. Having access to the information, but also being able to capitalize on that value that they are generating. Open source as well. Being able to contribute to the community from an open source standpoint. And we’re also supporting them to build their own data sets that are truly representative of their local context. Following with education. We’ve made a partnership with Zindia Africa where we’ve trained developers on our tools. Starting from helping them understand how to handle geospatial data, which is a key type of data for our countries. Which allows them to really understand where they are beginning the transition to really model the fundamental knowledge of everyone. We are really putting this knowledge into what we are building. And we are moving from just consuming technology to really shaping the future of these kids. Moving into efficiency, you know, if they told us something, we are forgotten to chase for efficiency. We are moving from really exploring global models to building local models, smaller models, more efficient models that are truly adapted to our context. From Kenya to Barbados to Nepal. So how do we win? We need to resist, we need to innovate, and we need to apply. We have to embrace global innovation. There is no way around it. But we have to do it in a way that is innovative and does stands who is I how it isahaha where we have to embrace innovation, but we have to demand that critical AI solutions serve our purposes and we have to reject them when they don’t and focus on strengthening our own innovation. And a perfect example of this is Barbados, where today, we are supporting the government to create a local R&D hub, equipping them with our first microdata center powered by renewable energy. Because not all our countries have to buy hundreds of billions of dollars of infrastructure. They need the right infrastructure in the right hands and with the right resources. We have to collaborate. None of this will be possible if we don’t build a collaborative framework. South to south collaboration. We have to understand and identify the countries who build, who produce, and who apply AI. This collaborative framework will ask and demand for an open dialogue. For knowledge sharing and mutual support. It’s a blueprint, a true blueprint for AI sovereignty that the rest of the world will learn from. A data-rich, community-driven, and including AI ecosystem will benefit not just our communities, but the entire world. You know, I always get asked this question. What will it take for Africa and the global south to catch up with AI? And we have to deconstruct and reconstruct how we’ve been thinking about this race. The question was never whether we would catch up. It was always whether we will be able to use AI as a true equalizer. For our communities and our countries. I truly believe that the AI revolution is not being written in Silicon Valley. I truly believe that the AI revolution is not being written in Silicon Valley. I truly believe that the AI revolution, in reality, it’s coming from Manila, from Bridgetown, from Nairobi, from Accra, from Jakarta, and from Bangui. And I truly believe that what our grandparents were dreaming is what is happening today. What they’ve been fighting for, for centuries, is our time to make happen. This is our moment. This is our revolution, and this is our own AI future. Thank you.


K

Kate Kallot

Speech speed

148 words per minute

Speech length

2044 words

Speech time

828 seconds

Current AI systems are designed to exclude entire regions, particularly the Global South, from participating in the AI future

Explanation

Kate argues that entire regions are at risk of being left out of the AI future not by accident, but by design. She emphasizes that while AI has become the default interface for millions, billions are still at risk of exclusion from this technological advancement.


Evidence

She states that ‘Where AI has become the default interface for millions, billions are still at risk of being left out’ and frames this as happening ‘by design’ rather than accidentally.


Major discussion point

AI exclusion of Global South


Topics

Development | Digital access | Human rights


There exists a massive global connectivity divide with nearly half the world’s population lacking internet access

Explanation

Kate highlights the fundamental infrastructure gap where close to half of the world’s population still lacks internet access. This connectivity divide serves as a barrier to AI participation and representation.


Evidence

She mentions that ‘close to half of the world’s population still do not have access to Internet’ and contrasts this with ‘500 million of users on the most popular AI chat bot today’ against ‘2.6 billion people that are unconnected.’


Major discussion point

Global connectivity divide


Topics

Development | Digital access | Infrastructure


A compute divide exists where 90% of global data center capacity is concentrated in the US and China, while Africa has less than 1% despite representing 19% of world population

Explanation

Kate presents stark statistics showing the massive imbalance in computational infrastructure globally. She demonstrates how the infrastructure needed to train AI systems is concentrated in just two countries, leaving the rest of the world, particularly Africa, severely underrepresented.


Evidence

She provides specific statistics: ‘90% of the world’s data center capacity still sits in two countries, US and China’ while ‘those two countries only contribute to 21.6% of the population.’ In contrast, ‘Africa, 19% of the world’s population, less than 1% of the world’s data center capacity.’


Major discussion point

Compute infrastructure inequality


Topics

Development | Infrastructure | Digital access


AI systems trained on internet data don’t represent the 2.6 billion unconnected people, making them invisible to these systems

Explanation

Kate argues that AI models trained on internet data inherently exclude the perspectives, knowledge systems, cultures, and languages of the billions of people who are not connected to the internet. This creates a fundamental representation problem where these populations become invisible to AI systems.


Evidence

She explains that ‘while those models are using data that’s on the Internet to train the systems, and you have half of the world’s population that is not on the Internet, who does it actually represent? It doesn’t represent our systems, our knowledge systems, our culture, our languages.’


Major discussion point

AI representation gap


Topics

Human rights | Cultural diversity | Development


Personal family history demonstrates how talented individuals from resource-rich countries have long questioned why their nations don’t develop despite having natural resources and talent

Explanation

Kate uses her grandfather Joseph’s story to illustrate the historical pattern of talented individuals from resource-rich countries questioning why their nations don’t develop. Her grandfather was a highly educated African who traveled the world in the 1950s but was troubled by his country’s lack of development despite its wealth in resources and talent.


Evidence

She describes her grandfather Joseph who ‘was one of the first African graduates of Saint-Cyr, the national police school in France’ and ‘went on to studying at Indiana Police Academy, at Washington Police Academy, and even met Kennedy’ but was bothered by the question of ‘How could our country, who is extremely rich in natural resources, extremely rich in talent, doesn’t actually develop itself like the places that he has visited?’


Major discussion point

Historical development challenges


Topics

Development | Capacity development


The legacy of colonial oppression and the fight for justice provides motivation for ensuring AI serves all populations

Explanation

Kate draws on her family’s experience with colonial oppression and resistance to frame the current AI inequality as a continuation of historical injustices. Her grandfather’s murder by a dictator and her grandmother’s imprisonment while continuing to educate her children serves as inspiration for fighting current technological exclusion.


Evidence

She recounts how her grandfather ‘was killed by the dictator of the day, named Bokassa’ and her grandmother ‘was put in surveillance residence, aka prison’ but ‘decided to hide learning textbooks into cassava bags so she could actually continue giving her kids an education.’ She states ‘My grandparents taught me the meaning of justice for the first time.’


Major discussion point

Historical justice and AI equity


Topics

Human rights | Development | Cultural diversity


Constraint-driven creativity has always been a superpower for the Global South, forcing different thinking approaches

Explanation

Kate argues that constraints have historically been a source of strength and innovation for the Global South, forcing creative solutions and different approaches to problems. She positions this as an advantage that can be leveraged in the AI revolution.


Evidence

She states ‘Constraints has always been our superpower. Specifically, constraint-driven creativity. It has always forced us to think differently. It has always forced us to go against the tides.’


Major discussion point

Innovation through constraints


Topics

Development | Capacity development | Interdisciplinary approaches


Historical examples like M-Pesa demonstrate how locally-developed solutions can be more successful than global alternatives

Explanation

Kate uses M-Pesa as an example of how locally-developed technology solutions can be more successful and innovative than global alternatives. M-Pesa was created to solve a specific local problem and achieved massive adoption, predating similar global solutions.


Evidence

She mentions ‘M-Pesa, who was created in 2007, long before the WeChat pay, the Apple pay of this world. And M-Pesa was created to solve one specific problem. The lack of banking infrastructure in Kenya. And today it’s being used by more than 96% of the population.’ She also references ‘William, the boy who harnessed the wind, who built a windmill to power his family home in Malawi during a drought.’


Major discussion point

Local innovation success


Topics

Development | Inclusive finance | Digital business models


Local talent like Clinton, a data scientist from Nairobi, shows the potential when people are empowered to solve problems in their communities

Explanation

Kate highlights Clinton as an example of local talent who, when empowered, can create innovative solutions for community problems. She uses his story to demonstrate the potential that exists when people are given the tools and opportunity to lead technological development in their regions.


Evidence

She describes ‘Clinton is a brilliant data scientist from Nairobi. And long before he actually graduated, he had built already three startups to solve really from the University of Michigan. He is a professor of computer science and he is a leader in solving problems that are around him for his communities. Think about deforestation, think about monitoring or wildlife conservation.’ She notes that ‘Clinton today is the head of data science at Amini and is actually sitting right here with you in the audience.’


Major discussion point

Empowering local talent


Topics

Development | Capacity development | Human rights


Amini is building data backbone infrastructure for the Global South as a regenerative rather than extractive layer

Explanation

Kate explains that Amini is creating infrastructure that benefits the communities generating the data rather than simply extracting value from them. This approach focuses on building regenerative systems that serve local populations while addressing data sovereignty and creating localized intelligence.


Evidence

She states ‘We’re building the data backbone for the global south. Not just as an extractive layer, but actually as a regenerative one. From the connectivity, the infrastructure, and the computes.’ She mentions they are ‘Addressing the questions of data sovereignty and residency. Of hyperlocalized intelligence. Sovereign compute infrastructure.’


Major discussion point

Building regenerative AI infrastructure


Topics

Development | Infrastructure | Data governance


The focus should be on building appropriate technology infrastructure that enables local talent to prosper, not chasing the largest models

Explanation

Kate argues that success won’t come from competing with the largest AI models or supercomputers, but from building the right infrastructure that empowers local talent. She emphasizes that building critical infrastructure today means creating technology infrastructure for local innovators.


Evidence

She states ‘we’re not going to win by chasing the largest model or the largest supercomputer. But we will win by harnessing our digitally native use and also supporting our communities.’ She explains that ‘building a critical infrastructure meant building a road, building a hospital, building a bridge. Today, it actually means building the technology infrastructure for the Clintons of your country to be able to prosper.’


Major discussion point

Appropriate technology infrastructure


Topics

Development | Infrastructure | Capacity development


Partnerships with local universities, governments, and communities are essential for creating hyperlocalized intelligence solutions

Explanation

Kate emphasizes the importance of working with local institutions and communities at every level to create AI solutions that are truly representative of local contexts. This collaborative approach ensures that the technology serves the communities it’s built for.


Evidence

She mentions ‘working with local universities, governments, and communities, and engaging them at every single layer of the ecosystem.’ She provides examples like ‘Starting with farmers in Nepal, we’re supporting farmers to digitize themselves. Enabling them to build data sets that are truly representative of their local context.’


Major discussion point

Community-centered AI development


Topics

Development | Capacity development | Cultural diversity


The approach must involve resisting inappropriate solutions, innovating locally, and applying global innovations selectively

Explanation

Kate outlines a three-pronged strategy for AI sovereignty that involves critically evaluating global solutions, demanding that AI serves local purposes, and focusing on strengthening local innovation capabilities. This approach balances global engagement with local empowerment.


Evidence

She states ‘We need to resist, we need to innovate, and we need to apply. We have to embrace global innovation. There is no way around it. But we have to do it in a way that is innovative and does stands who is I how it is… we have to demand that critical AI solutions serve our purposes and we have to reject them when they don’t and focus on strengthening our own innovation.’


Major discussion point

Strategy for AI sovereignty


Topics

Development | Human rights | Data governance


South-to-south collaboration is crucial for building a framework that supports countries in producing and applying AI

Explanation

Kate advocates for collaboration between Global South countries to create a framework for AI development that includes knowledge sharing, mutual support, and coordinated efforts. This collaboration would create a blueprint for AI sovereignty that could benefit the entire world.


Evidence

She emphasizes ‘We have to collaborate. None of this will be possible if we don’t build a collaborative framework. South to south collaboration. We have to understand and identify the countries who build, who produce, and who apply AI.’ She describes this as ‘a blueprint, a true blueprint for AI sovereignty that the rest of the world will learn from.’


Major discussion point

South-to-south collaboration


Topics

Development | Capacity development | Data governance


The AI revolution is actually emerging from cities like Manila, Nairobi, and Jakarta rather than Silicon Valley

Explanation

Kate challenges the conventional narrative that AI innovation is centered in Silicon Valley, arguing instead that the real AI revolution is happening in Global South cities. She positions this as the fulfillment of what previous generations fought for and dreamed of achieving.


Evidence

She declares ‘I truly believe that the AI revolution is not being written in Silicon Valley. I truly believe that the AI revolution, in reality, it’s coming from Manila, from Bridgetown, from Nairobi, from Accra, from Jakarta, and from Bangui.’ She connects this to her family legacy: ‘what our grandparents were dreaming is what is happening today.’


Major discussion point

Global South as AI innovation center


Topics

Development | Capacity development | Cultural diversity


M

Moderator

Speech speed

76 words per minute

Speech length

37 words

Speech time

29 seconds

Kate Kallot is introduced as the founder and CEO of Amini with an enthusiastic welcome to the stage

Explanation

The moderator provides a warm and enthusiastic introduction for Kate Kallot, highlighting her role as founder and CEO of Amini and expressing excitement about having her speak. The introduction sets a positive tone for the presentation.


Evidence

The moderator states ‘founder and CEO at Amini. It’s absolutely fabulous to have her on the stage. I’ve seen her speak before. Please join me in welcoming a pre-lunch welcome, please, with slightly more coffee. It’s the fabulous Kate Kallot.’


Major discussion point

Speaker introduction


Agreements

Agreement points

Similar viewpoints

Unexpected consensus

Overall assessment

Summary

This transcript represents a single-speaker presentation rather than a multi-speaker discussion or debate. Kate Kallot delivered a comprehensive speech about AI inequality and the need for Global South empowerment, with only a brief introductory comment from the moderator.


Consensus level

No consensus analysis is possible as there was only one substantive speaker presenting arguments. The moderator’s role was purely introductory and ceremonial, expressing enthusiasm for Kate Kallot’s participation but not engaging with or responding to any of her arguments about AI sovereignty, digital divides, or Global South empowerment. This was a keynote-style presentation rather than a collaborative discussion where agreement or disagreement could emerge between multiple participants.


Differences

Different viewpoints

Unexpected differences

Overall assessment

Summary

No disagreements identified as this is a single-speaker presentation rather than a multi-speaker debate or discussion


Disagreement level

No disagreement present – Kate Kallot delivers a cohesive presentation about AI inequality and her company’s solutions, with only a supportive introduction from the moderator. The format is a keynote speech rather than a debate, so there are no opposing viewpoints or conflicting arguments presented.


Partial agreements

Partial agreements

Similar viewpoints

Takeaways

Key takeaways

Current AI systems are systematically excluding the Global South by design, creating a digital divide where billions remain invisible to AI systems trained on internet data


The Global South faces three critical divides: connectivity (half the world lacks internet), compute infrastructure (90% of data centers in US/China vs <1% in Africa), and representation in AI systems


Constraint-driven creativity and local innovation have historically been superpowers for the Global South, as demonstrated by solutions like M-Pesa


Building critical AI infrastructure means empowering local talent and communities rather than chasing the largest models or supercomputers


AI sovereignty requires a three-pronged approach: resist inappropriate solutions, innovate locally, and selectively apply global innovations


South-to-south collaboration is essential for creating a framework where Global South countries can produce and apply AI rather than just consume it


The true AI revolution is emerging from cities like Manila, Nairobi, and Jakarta, not Silicon Valley


Technology must be invisible and serve local communities to be truly useful, focusing on regenerative rather than extractive approaches


Resolutions and action items

Build data backbone infrastructure for the Global South through Amini’s platform


Create partnerships with local universities, governments, and communities for hyperlocalized intelligence solutions


Train developers on geospatial data handling and AI tools through partnerships like Zindia Africa


Support farmers in Nepal and other regions to digitize and build representative datasets while retaining value


Establish local R&D hubs with microdata centers powered by renewable energy, starting with Barbados


Develop smaller, more efficient models adapted to local contexts from Kenya to Barbados to Nepal


Focus on building appropriate technology infrastructure that enables local talent to prosper


Unresolved issues

How to scale the local-first approach across all Global South regions effectively


Specific mechanisms for ensuring data sovereignty and preventing extraction by global tech companies


Detailed framework for south-to-south collaboration and knowledge sharing


How to balance embracing global innovation while maintaining local control and benefits


Funding and resource allocation strategies for building distributed AI infrastructure


Regulatory and policy frameworks needed to support AI sovereignty initiatives


Suggested compromises

Embrace global innovation selectively while demanding that AI solutions serve local purposes


Build collaborative frameworks that allow for knowledge sharing while maintaining sovereignty


Focus on appropriate-scale infrastructure rather than competing with massive global data centers


Combine open source contributions with community-driven development approaches


Thought provoking comments

We are living through a time where entire regions are at risk of being left out of the future. And that’s not happening by accident. It’s actually happening by design.

Speaker

Kate Kallot


Reason

This opening statement is profoundly insightful because it reframes the AI divide from an inevitable consequence of technological progress to a deliberate structural choice. By using the word ‘design,’ Kallot challenges the audience to see AI inequality not as natural or accidental, but as the result of conscious decisions that can therefore be changed.


Impact

This comment sets the entire tone and framework for the presentation, establishing that the discussion will challenge conventional narratives about technological progress and examine power structures. It immediately positions the audience to think critically about agency and responsibility in AI development.


While those models are using data that’s on the Internet to train the systems, and you have half of the world’s population that is not on the Internet, who does it actually represent? It doesn’t represent our systems, our knowledge systems, our culture, our languages. It doesn’t represent us. We invisible to those systems.

Speaker

Kate Kallot


Reason

This observation is deeply thought-provoking because it exposes a fundamental flaw in how we conceptualize AI training data as ‘universal.’ Kallot reveals that what we consider comprehensive global knowledge is actually a narrow slice representing only the connected half of humanity, making AI systems inherently biased and incomplete.


Impact

This comment shifts the discussion from technical infrastructure gaps to epistemological exclusion – the idea that entire ways of knowing and being are being erased from AI systems. It deepens the conversation by moving beyond access issues to questions of representation and cultural preservation.


Today we’re standing on the edge. We’re standing on the edge where AI can be the final pin to erasing centuries of our knowledge systems, of our culture, of our languages. Can deepen the extraction that we have seen in our countries. Or can be the most inclusive leap in history.

Speaker

Kate Kallot


Reason

This is a pivotal insight that frames AI as an existential inflection point for the Global South. Kallot presents AI not as neutral technology but as having the potential for either cultural genocide or unprecedented empowerment, forcing the audience to confront the stakes involved.


Impact

This comment creates a crucial turning point in the presentation, moving from problem identification to urgency for action. It elevates the discussion from technical concerns to civilizational ones, making the case that this is a defining moment requiring immediate response.


Technology, to truly be useful, it has to be invisible.

Speaker

Kate Kallot


Reason

This deceptively simple statement contains profound wisdom about technology adoption and design. Drawing from the M-Pesa example, Kallot suggests that the most transformative technologies are those that seamlessly integrate into existing social and economic patterns rather than requiring users to adapt to them.


Impact

This insight redirects the conversation from flashy, complex AI systems to practical, contextually appropriate solutions. It provides a design philosophy that challenges the prevailing narrative of AI as necessarily complex and visible, suggesting instead that true innovation lies in simplicity and integration.


Constraints has always been our superpower. Specifically, constraint-driven creativity. It has always forced us to think differently. It has always forced us to go against the tides.

Speaker

Kate Kallot


Reason

This comment is transformative because it reframes limitations as advantages. Instead of seeing resource constraints as obstacles to overcome, Kallot positions them as catalysts for innovation that can lead to superior solutions. This challenges deficit-based thinking about the Global South.


Impact

This insight fundamentally shifts the narrative from one of catching up to one of leading through different approaches. It empowers the audience to see their constraints not as disadvantages but as sources of competitive advantage and unique innovation pathways.


The question was never whether we would catch up. It was always whether we will be able to use AI as a true equalizer.

Speaker

Kate Kallot


Reason

This statement deconstructs the entire framing of technological development as a race to be won or lost. Kallot challenges the assumption that the Global South should aspire to replicate existing AI development patterns, instead proposing that the goal should be leveraging AI for equity and empowerment.


Impact

This comment provides a powerful reframe for the entire discussion, moving away from competitive metaphors toward collaborative and equity-focused ones. It suggests that success should be measured not by technological parity but by social impact and empowerment.


Overall assessment

These key comments collectively transformed what could have been a typical technology presentation into a profound examination of power, representation, and agency in the AI era. Kallot’s insights consistently challenged dominant narratives about technological progress, reframing constraints as advantages, problems as opportunities, and the Global South’s role from passive recipient to active creator. The comments built upon each other to create a compelling argument for AI sovereignty that goes beyond technical solutions to address fundamental questions of cultural preservation, economic justice, and self-determination. The presentation’s power lies in how these insights work together to completely reframe the AI development conversation from one of technological catch-up to one of indigenous innovation and community empowerment.


Follow-up questions

What does it take for the global South to not just be connected to an AI future, to not just be carried into an AI future, but actually authoring its own AI future?

Speaker

Kate Kallot


Explanation

This is the foundational question that motivated Kate to start Amini and represents the core challenge her work addresses – moving from passive participation to active leadership in AI development


Whose freedom is today’s AI actually serving?

Speaker

Kate Kallot


Explanation

This question challenges the current AI paradigm and highlights issues of representation, access, and benefit distribution in AI systems globally


How do we create more Clintons? How do we empower them to lead? How do we give them the tools to build the future of their communities and their countries?

Speaker

Kate Kallot


Explanation

These questions focus on scaling talent development and empowerment in the global South, using Clinton as an example of local innovation potential


How do we turn the tides today?

Speaker

Kate Kallot


Explanation

This question addresses the strategic approach needed to shift from the current extractive AI model to one that empowers the global South


What will it take for Africa and the global south to catch up with AI?

Speaker

Kate Kallot


Explanation

Kate mentions this as a frequently asked question that needs to be deconstructed and reframed, suggesting the need for research into alternative approaches to AI development that don’t assume a ‘catching up’ model


Research into addressing data sovereignty and residency challenges

Speaker

Kate Kallot


Explanation

This area requires further investigation to ensure that data generated in the global South benefits local communities rather than being extracted for external use


Research into hyperlocalized intelligence development

Speaker

Kate Kallot


Explanation

This involves studying how to create AI systems that are specifically adapted to local contexts, languages, and knowledge systems


Research into sovereign compute infrastructure models

Speaker

Kate Kallot


Explanation

This requires investigation into how countries in the global South can develop their own computing infrastructure without requiring massive investments like those made by the US and China


Research into South-to-South collaboration frameworks for AI

Speaker

Kate Kallot


Explanation

This involves studying how countries in the global South can collaborate with each other to build AI capabilities, rather than relying solely on technology transfer from developed nations


Disclaimer: This is not an official session record. DiploAI generates these resources from audiovisual recordings, and they are presented as-is, including potential errors. Due to logistical challenges, such as discrepancies in audio/video or transcripts, names may be misspelled. We strive for accuracy to the best of our ability.