Global Digital Compact: AI solutions for a digital economy inclusive and beneficial for all

Summary

This discussion focused on AI solutions for creating an inclusive and beneficial digital economy, held as part of the World Summit on the Information Society (WSIS) Forum 2025. The session was moderated by Isabel de Sola from the UN Office for Digital and Emerging Technologies and featured speakers from various UN agencies and private sector organizations. The conversation began by acknowledging that while we are already living in a digital economy, it has not achieved the inclusive benefits originally envisioned 20 years ago when the WSIS framework was established.

Key challenges identified include a critical shortage of AI-skilled workers, with 85 million jobs potentially unfilled by 2030, lack of reliable data and infrastructure, and limited purchasing power for AI technologies among smaller businesses. Siyong Zou from UNIDO emphasized that AI transformation is inevitable, but the focus must be on ensuring it serves everyone, particularly those historically left behind by technological progress. Amandeep Singh Gill, the UN Secretary General’s Special Envoy on Technology, highlighted how the Global Digital Compact places the inclusive digital economy at its center and stressed the importance of building comprehensive digital ecosystems rather than pursuing isolated AI solutions.

The session featured several concrete AI solution examples from around the world. These included CGI’s environmental monitoring system using satellite data and AI to predict water pollution, Beijing Institute of Technology’s AI Green Index for measuring sustainable AI practices, Zindi’s platform connecting 90,000 data scientists globally through competitive challenges, and humanitarian AI tools from Data Friendly Space that provide rapid crisis response analysis. Microsoft announced its new Elevate program, committing $4 billion over five years for AI education and technology donations, while NTT Data showcased AI applications for workplace safety and accessibility. The session concluded with the launch of a global call for solutions to foster digital inclusion, seeking innovative approaches to empower marginalized communities and build a more equitable digital economy.

Keypoints

## Major Discussion Points:

– **Digital Economy Inclusion Challenges**: The discussion highlighted significant barriers to an inclusive digital economy, including the fact that 80% of companies worldwide lack web pages, there’s a critical shortage of AI-qualified workers (85 million jobs may go unfilled by 2030), and persistent digital divides in infrastructure, data access, and purchasing power for AI technologies.

– **Global Digital Compact Framework**: The newly approved Global Digital Compact (September 2024) was presented as a key framework for addressing digital economy inclusion, with Objective 2 specifically focused on ensuring no worker, enterprise, or country is left behind in digital transformation.

– **AI Solutions Showcase**: Multiple organizations presented concrete AI solutions addressing various aspects of inclusion: environmental monitoring and pollution prediction (CGI), AI sustainability measurement (AI Green Index), youth skill development through competitions (Zindi), humanitarian crisis response (Data Friendly Space/AWS), and workforce empowerment across industries (NTT Data).

– **Ecosystem Approach to Digital Development**: Speakers emphasized that successful AI deployment requires building comprehensive digital ecosystems rather than isolated solutions, including digital public infrastructure, policy frameworks, skills development, and public-private partnerships.

– **Global Call for Solutions Launch**: The session concluded with the announcement of a new global initiative calling for innovative digital solutions to empower underserved communities, with applications open to innovators, entrepreneurs, and organizations worldwide.

## Overall Purpose:

The discussion aimed to assess the current state of digital economy inclusion 20 years after the WSIS framework, identify barriers preventing equitable access to AI and digital technologies, showcase practical AI solutions addressing these challenges, and launch a global call for innovative solutions to foster a more inclusive digital economy aligned with the UN’s 2030 Sustainable Development Goals.

## Overall Tone:

The discussion maintained a consistently optimistic and action-oriented tone throughout. While speakers acknowledged serious challenges and the widening digital divide, the emphasis remained on solutions, collaboration, and concrete examples of successful AI implementations. The tone was professional yet energetic (as noted by the moderator’s reference to “high energy” style), with speakers demonstrating enthusiasm for their initiatives and a shared commitment to leaving no one behind in the digital transformation. The announcement of the global call for solutions at the end reinforced the forward-looking, collaborative spirit of the entire session.

Speakers

– **Isabel de Sola** – UN Office for Digital and Emerging Technologies, Moderator

– **Ciyong Zou** – Deputy to the Director General of UNIDO

– **Amandeep Singh Gill** – UN Secretary General’s Special Envoy on Technology

– **Mattie Yeta** – CGI (UK), presented via video

– **LIU Hao** – Beijing Institute of Technology (BIT), China, presented AI Green Index

– **Celina Lee** – CEO and co-founder of Zindi, AI innovation in Africa

– **Sasha Rubel** – AWS

– **Doug Smith** – Acting CEO of Data Friendly Space

– **Jean‑Francois Saint‑Pierre** – Microsoft, formerly with tech for social impact team

– **Manel Martorana** – NTT Data, IT services provider

– **Jason Slater** – Role/title not specified, involved in closing remarks and global call announcement

**Additional speakers:**

None identified beyond the provided speakers names list.

# AI Solutions for Creating an Inclusive and Beneficial Digital Economy: Comprehensive Discussion Report

## Executive Summary

This high-energy, rapid-fire discussion, conducted in “Jason Slater style” as part of the World Summit on the Information Society (WSIS) Forum 2025, brought together representatives from UN agencies, technology companies, and research institutions to address the critical challenge of creating an inclusive digital economy through artificial intelligence solutions. Moderated by Isabel de Sola from the UN Office for Digital and Emerging Technologies, the 45-minute post-lunch session examined how AI can bridge existing digital divides whilst ensuring that technological advancement serves all populations.

The conversation revealed a stark reality: whilst we are already living in a digital economy, it has failed to achieve the inclusive benefits originally envisioned when the WSIS framework was established two decades ago. With significant disparities in digital access and AI adoption across countries and industries, the digital divide continues to widen rather than narrow. The session concluded with the announcement of a Global Initiative for Digital Inclusion, representing a new global call for solutions to foster digital inclusion across five thematic areas.

The discussion demonstrated remarkable consensus among speakers on the need for ecosystem approaches to AI deployment, human-centred implementation, and the urgency of addressing digital exclusion through collaborative action between public and private sectors.

## Setting the Context: Twenty Years of WSIS and Current Challenges

Isabel de Sola opened the session by reflecting on the twenty years of engagement since the WSIS framework was established, noting that whilst we now live in a digital economy, it lacks the inclusivity that was originally hoped for. The statistics compiled by her office paint a concerning picture of persistent digital divides and uneven AI uptake across countries and industries.

The Global Digital Compact, approved in September 2024, was presented by Amandeep Singh Gill (who arrived late from an “AI governance lunch”) as a crucial framework for addressing these challenges. Unlike previous initiatives that focused primarily on principles, the Compact provides an action-oriented agenda with the inclusive digital economy at its centre. Objective 2 of the Compact specifically focuses on ensuring that no worker, enterprise, or country is left behind in digital transformation.

Gill raised a fundamental question about measurement: how do we define and measure the digital economy when global estimates vary widely between 10-40% of global GDP? This uncertainty about what constitutes the digital economy complicates efforts to develop targeted policies and measure progress towards inclusion.

## The Urgency of Action: Skills Gaps and Infrastructure Deficits

Ciyong Zou from UNIDO provided perhaps the most alarming statistic of the discussion: the global digital skills shortage threatens to leave 85 million jobs unfilled by 2030. As Zou emphasised, “Without deliberate targeted action, the digital divide will not just persist, it will deepen, creating a world where the benefits of AI are concentrated among the very few, while the many are left further behind.”

The infrastructure challenges are equally stark. Mattie Yeta from CGI, presenting via video, highlighted the disparity in internet access, with 93% usage in high-income countries compared to just 27% in low-income countries, leaving 2.6 billion people without internet access. This digital divide extends beyond mere connectivity to encompass data accessibility, reliable infrastructure, and purchasing power for AI technologies.

## The Ecosystem Approach: Building Fertile Ground for AI

A central theme throughout the discussion was the need for comprehensive ecosystem approaches rather than isolated technological solutions. Amandeep Singh Gill provided crucial insight into successful AI deployment: “If we ask the questions: why are the US and China so successful in deploying AI solutions? It’s because there is a broad base of the digital economy. So those seeds are falling on fertile land.”

This ecosystem thinking challenges the common approach of attempting to leapfrog directly to advanced AI without building foundational digital infrastructure. Gill emphasised that countries need to take a strategic approach to building the digital economy rather than engaging in “a blind race to match OpenAI, DeepSeek.” Successful AI implementation requires foundational elements including digital public infrastructure, policy frameworks, cybersecurity, e-commerce capabilities, and government services.

Ciyong Zou reinforced this perspective through UNIDO’s Global Alliance approach, which focuses on building scalable AI platforms through collaboration with 120 partners across 40 countries. Rather than pursuing isolated pilot projects, UNIDO emphasises creating comprehensive industrial ecosystems that make AI affordable and actionable for local industries.

## Concrete AI Solutions: Demonstrations of Impact

The session featured several compelling examples of AI solutions addressing various aspects of digital inclusion, demonstrating how AI can be deployed effectively when properly integrated into broader digital ecosystems.

### Environmental Monitoring and Sustainability

Mattie Yeta presented CGI’s environmental monitoring system, which uses satellite data and AI to predict water pollution. The system combines datasets from earth observation, sensors, and weather patterns to identify pollution from multiple sources and predict pollution likelihood.

Liu Hao from Beijing Institute of Technology introduced the AI Green Index, a joint product between UNIDO and Beijing Institute of Technology. This comprehensive measurement system features five pillars and 18 indicators designed to make AI environmentally sustainable. The index addresses the dual challenge of using AI to solve environmental problems whilst ensuring that AI itself operates sustainably, providing a framework for measuring and regulating AI’s environmental impact.

### Skills Development and Capacity Building

Celina Lee from Zindi presented a compelling model for AI skills development through her platform connecting 90,000 data scientists across 180 countries. The platform uses competitive challenges to build real-world AI skills. In Kenya, with 9,000 users and 2,000 achieving job outcomes, Zindi found that participants who entered five competitions and joined teams achieved positive job outcomes regardless of whether they won. This insight demonstrates that participation and collaboration matter more than individual excellence.

Jean-François Saint-Pierre from Microsoft announced the Elevate programme just 24 hours before the session (at 6 p.m. the previous day), representing a significant commitment for AI education and technology donations. The Microsoft Elevate Academy aims to help 20 million people learn AI skills over the next two years, addressing the massive skills gap through scalable, accessible training programmes.

### Humanitarian and Crisis Response

Doug Smith from Data Friendly Space presented AI-enabled humanitarian response tools that can provide rapid analysis within three hours of disasters. Using a hypothetical example of a Myanmar earthquake (which he clarified was not a real event), Smith demonstrated how the Gannet system’s three main components—virtual assistant, situation hub, and private workspace—allow conversational interaction with data and provide regular updates during crises.

Sasha Rubel from AWS emphasised the importance of reducing “the space between data, information and insights and life-saving action,” redefining AI not merely as a data processing tool but as a bridge to actionable intervention in humanitarian contexts.

### Industrial and Workplace Applications

Manel Martorana from NTT Data showcased AI applications addressing workplace challenges including worker shortages, accessibility for people with disabilities, and workplace safety. These applications demonstrate how AI can augment human capabilities rather than replace workers, making technology more accessible and affordable across various industrial settings.

## Corporate Responsibility and Partnership Models

The discussion revealed strong alignment among major technology companies on using their resources and platforms to democratise AI access rather than focusing primarily on commercial applications. Microsoft’s Elevate commitment, AWS’s humanitarian partnerships, and NTT Data’s focus on accessibility all demonstrate how private sector organisations can contribute to inclusive digital economy development.

This alignment between UN agencies and private sector organisations suggests significant potential for coordinated global action, with speakers consistently emphasising human-centred AI implementation and the importance of human-in-the-loop systems designed to augment rather than replace human capabilities.

## The Global Call for Solutions: A New Initiative

The session concluded with Jason Slater announcing the launch of the Global Initiative for Digital Inclusion, representing a new global call for solutions to foster digital inclusion. This initiative seeks innovative approaches across five thematic areas: skills empowerment, enabling policy, innovation acceleration, sustainable digital supply chains, and additional areas to be specified.

Selected solutions will be promoted during the UN General Assembly’s 80th session and UNIDO’s general conference, providing a pathway for scaling successful innovations. The call for solutions represents a concrete step towards implementing the collaborative, ecosystem-based approaches discussed throughout the session.

## Challenges and Unresolved Issues

Despite the optimistic tone and concrete solutions presented, several significant challenges remain unresolved. The measurement problem identified by Amandeep Singh Gill represents a fundamental challenge: how can we build an inclusive digital economy without clear metrics for what constitutes success?

The standardisation of AI governance frameworks, particularly for environmental sustainability, remains incomplete. Liu Hao noted that whilst Europe has some AI regulations, other regions lack adequate frameworks for measuring and regulating AI’s environmental impact.

The challenge of scaling successful pilot projects to comprehensive industrial ecosystems persists. Whilst several speakers presented successful local implementations, the question of how to adapt and scale these solutions across different contexts and regions requires continued attention.

## Key Agreements and Consensus

The discussion demonstrated remarkable consensus among speakers on fundamental challenges and goals. All participants agreed on the urgency of addressing digital exclusion, the critical nature of skills gaps, and the need for ecosystem approaches to AI deployment. The alignment between UN agencies and private sector organisations on human-centred AI implementation suggests a mature understanding of AI’s role in inclusive development.

Speakers consistently emphasised that successful AI deployment cannot be achieved through isolated technological solutions but requires comprehensive digital transformation addressing infrastructure, skills, policy frameworks, and sustainable implementation practices simultaneously.

## Conclusion

This rapid-fire discussion highlighted both the urgency and the opportunity inherent in creating an inclusive digital economy through AI. The 85 million jobs that may go unfilled by 2030 and the 2.6 billion people without internet access represent significant challenges, but the solutions presented demonstrate that progress is possible when AI is deployed within comprehensive digital ecosystems.

The consensus among speakers on the need for collaborative, human-centred approaches to AI development provides a foundation for coordinated action. The combination of UN framework initiatives like the Global Digital Compact, substantial private sector commitments, and the new Global Initiative for Digital Inclusion suggests potential for more coordinated global action in the next phase of digital economy development.

However, success will require sustained commitment to building foundational digital infrastructure, addressing skills gaps through innovative training approaches, and ensuring that AI solutions remain accessible and relevant to the communities they are meant to serve. The global call for solutions represents an important step towards mobilising the innovation and collaboration necessary to achieve these ambitious goals.

As the session demonstrated, whilst we are already living in a digital economy, the work of making it truly inclusive requires deliberate, targeted action across multiple dimensions. The AI solutions and approaches presented offer concrete examples of how this can be achieved when technology is deployed within comprehensive, human-centred frameworks designed to serve all populations.

Isabel de Sola: Good afternoon, everyone. Good morning. Good evening. Can you hear me okay online. Thank you. I am Isabel de Sola, I am part of the UN Office for Digital and Emerging Technologies, and I have the great pleasure to be your moderator for this exciting session, AI solutions for an inclusive and beneficial digital economy today. So we have 45 minutes and this is going to be in the style of Jason Slater and if you know him you know that it’s high energy. It’s after lunch. I hope all of you had a chance to have a coffee here in Geneva. So we have 45 minutes and what we’d like to do first is set the scene. What is even an inclusive and beneficial digital economy. And then we’d like to provide some examples concrete examples of AI solutions for organizations of all sizes that can showcase the different ways in which AI will power the digital economy or continue to transform the digital economy. I’d like to just take a note as well or share with all of you my immense pleasure that it’s been 20 years now that the WSIS framework has been engaging on the topic of the digital economy. So the WSIS level event 2025 is particularly important because it gives us a chance to look back. If you read the original texts of the WSIS on the digital economy it was quite hopeful. 20 years ago, these digital technologies were going to come into small, medium, and large organizations and help them buy and sell goods and services, help them find the clients, partners, supply parts that they needed, and everything was going to be great. I think part of our reflection today will be to take stock that we are in the digital economy already. However, it’s not as incredibly inclusive for all as we expected. So we now have some statistics that can describe that thanks to different UN offices, some of them who are in the room, UNCTAD that produces a report every year on the state of the digital economy. the International Trade Center, the WTO, they have noticed that there’s a certain trend towards large organizations in the digital economy. Using that instead of to say a trend towards concentration or monopolization in the digital economy. We know that 80% of companies worldwide don’t have web pages, for example, and the uptake of AI varies from country to country and from industry to industry. So with that as our backdrop, I’d like to introduce two very thoughtful speakers. The Deputy to the Director General of UNIDO, Mr. Siyong Su. Over to you for some scene-setting remarks.

Ciyong Zou: Thank you. Thank you very much, moderator. Distinguished representatives, ladies and gentlemen, good afternoon. A warm welcome to this pivotal session on AI solutions for an inclusive digital economy. We gather here today at a defining moment in human history, where artificial intelligence is not merely changing how we work. It is fundamentally reshaping the very fabric of industries, economies, and societies around the world. The question we face is not whether AI will transform our world, but rather how we can ensure that this transformation serves everyone, particularly those who have been historically left behind by technological progress. At UNIDO, we view AI as a powerful enabler of inclusive and sustainable industrial development through smart manufacturing, data-driven value chains, and intelligent production systems, supported by our network of centers of excellence. We have witnessed how AI can help developing countries leapfrog entire infrastructural stages and build resilient and climate-smart economies. However, we must confront the stark realities that threaten to undermine this potential. Perhaps most alarming is that our global digital skills shortage threatens to leave 85 million jobs unfilled by 2030. Without deliberate targeted action, the digital divide will not just persist, it will deepen, creating a world where the benefits of AI are concentrated among the very few, while the many are left further behind. This is precisely why UNIDO has taken on the responsibility of co-leading Objective 2 of the Global Digital Compact, placing the inclusive digital economy at the very heart of the international development agenda. Our mission is to ensure that no worker, no enterprise, and no country is left behind in this digital transformation. This is easy to talk about, but difficult to realize. We must work together to unlock AI’s potential to create jobs for everybody, every country. In this regard, UNIDO is now working together with partners from the public and private sectors of the UN system to drive this kind of inclusive application of AI in different sectors, particularly in the manufacturing sector. I think this is really an issue we have to address together. Distinguished representatives, today’s event represents more than just an event. important conversation. It is a call to action, a call to implementation. UNIDO is moving beyond the pilot project, face towards building scalable platforms and the comprehensive industrial AI ecosystems. Through our flagship in global alliance, we are collaborating with over 120 partners across 40 countries, including leading technology firms and the industrial players to make AI affordable, adaptable and actionable for local industries worldwide. We are not just talking about AI, we are building AI skill systems, accelerating investor matchmaking and transferring proven solutions directly to factories and supply chains, where they can create immediate impact. Our approach recognizes that for AI to truly serve in inclusive development, it must be grounded in the reality of local context, accessible to micro, small and medium enterprises, and designed to strengthen rather than displace home capabilities. I’m particularly excited about today’s rapid fire solution, which will showcase exactly the kinds of cost-effective, context-aware innovations we need to close the digital gaps and bring the transformative benefits of AI to the global source. These presentations will demonstrate how inclusive innovation can bridge the digital divide and create benefits that extend far beyond individual companies or countries. As we move forward, let us remember that AI’s true success lies not in advanced algorithm or faster processing, but in ensuring this digital revolution is guided by human value, inclusion, equity, and the real-world industrial impact. The future of AI must be a future where technology serves humanity, where digital transformation reinforces our commitment to leaving no one behind, and the benefits of innovation reach every corner of the world. I invite you all to join this mission. Together, let us make the AI revolution a force for equity, sustainable development, and shared prosperity. Thank you for your kind attention.

Isabel de Sola: Thank you, Mr. Zou, and as Dr. Amandeep Singh Gill, the UN Secretary General’s Special Envoy on Technology, has just joined us, I’ll give him a minute or so to gather his thoughts on how the GDC, the Global Digital Compact, which was approved in September of 2024, gives a framework within which we can work on the digital inclusive economy. Our office that supported the negotiation of the GDC conducted some research or gathered some data points that I’d like to share with you now about what are some of the hurdles to the digital inclusive economy according to small and medium-sized enterprises. We gathered this data from the ITC and from UNCTAD reports and from different sources to better understand what’s standing in the way of bringing in all shapes and sizes into this digital economy. Mr. Su mentioned some of them. Lack of qualified and certified workforce on AI-related jobs was the number one concern of the companies that participated in this research. Lack of reliable data as the fuel of AI that could be useful in specific contexts. Lack of reliable infrastructure, so the old digital divide, electricity, computing power, data centers, and also low purchasing power from businesses of different sizes to pay for the licenses of pre-trained frameworks. So those are some of the challenges that were in the backdrop of the negotiation of Objective 2 of the GDC. And I’m very pleased to invite Amandeep to share some remarks with us on that question.

Amandeep Singh Gill: Thank you very much. Thank you, Isabel. And good afternoon, everyone. DDG, good to be here with you. Apologies, the AI governance lunch is just about to finish. I managed to get away. Sorry for being late. Wherever you go these days, any region, when you meet with leaders, either from the public sector or the private sector, an inclusive digital economy is top of mind. Quality jobs and using the digital opportunity to leapfrog the development challenges. So that’s why it was natural for the Global Digital Compact to center the inclusive digital economy in terms of its five objectives, five recommendations. Some of them are enduring. They are a continuation of our agenda from the past. Some of them are emerging, AI governance, for example. But this is where the center really is, you know, center of gravity of the Global Digital Compact. And fortunately, we have an agenda, an action agenda. So it’s not just a description of the challenge. It’s not only principles and what needs to be done in analytical terms, but you know, actual actions. Some actions that have been tried over the past few years and have come out successfully, for example, investments in digital public infrastructure, where you can make every dollar count more than how ICT spend happens in the global north, for example. The points that the DDG was making in terms of, you know, harnessing some of these emerging technologies where the barriers to adoption may be lower. I mean, we still have to see, so question mark, fingers crossed. But I think with Gen-AI, we may see some barriers to adoption come down. Other barriers to development may actually go up, but some barriers to adoption may come down. So those are kind of action areas. In the work that we’ve done since then on the SG’s report on capacity building for AI, we’ve also looked at different maturity levels and what it needs, what they imply in terms of graduation pathways, countries going to higher levels of maturity on the digital economy. We’ve also looked at ways to measure the digital economy. Today, it’s a bit fuzzy. If you look at global estimates, anywhere between 10 to 20%, some people say that in a short while, this could be up to 40% of the global GDP. But how do we count this? What do we leave out? Is it only ICT spend? It’s something else. There’s some interesting research underway and we need to kind of prioritize that. And I’m so glad that UNIDO and UNCTAD are taking the lead in sub-objective two of the UN Systems Working Group on Digital Technologies to focus on some of these kind of issues. Lastly, just to close with a kind of a building block approach we need for this topic. So AI solutions, they could be discrete, but they best serve the purpose if they emerge from an organic ecosystem on the digital economy. If we ask the questions. You know, why are the US and China so successful in deploying AI solution? It’s because there is a broad base of the digital economy. So those seeds are falling on fertile land. So this is where I think countries, rather than, you know, a blind race to kind of match open AI, deep seek, you know, GPE for GPU need to take a strategic approach to building the digital economy. The foundation of digital public infrastructure, policy and regulation for data and skills, investments in cybersecurity, early use cases around e-commerce, government services, delivery of government services, where, you know, you can let the private sector take the lead. And there are some areas where public sector may take the lead, but other areas like health and agriculture where the two must combine forces. So then that kind of creates the ecosystem in which if you have sufficient data flows going through DPIs, you have all these, you know, ecosystems of collaboration coming up, then AI starts, AI solutions start to emerge in an organic way. So I think that is the approach that UNIDO and UNCTAD and all of us at the UN are trying to advocate and promote. Thank you.

Isabel de Sola: Thank you, Amandeep. That’s an excellent transition, actually, to begin looking at some of the AI solutions that will be presented now as inspirations, as examples, as catalysts for others. So we’re going to do a tour de force around the world of AI solutions that have grown up in their context, starting with Mattie Yeta from CGI in the UK who sent a video.

Mattie Yeta: CGI is excited to bring be part of the World Summit on the Information Society Forum 2025. We are one of the largest IT and business consulting services firms in the world. Our work with UNIDO is centered around developing innovative solutions such as AI, digital twins, blockchain, earth observations for sustainable development through our SEEDS program, Sustainability, Exploration and Environmental Data Science. Despite seeing massive ICT adoption in some places and the speed of innovations like AI in other places, there’s still a widening digital divide locally and globally. For example, internet usage in high-income countries sits at 93% whilst in low-income countries it sits at 27%. That means 2.6 billion people around the world are without internet. The adoption of digital technologies in countries around the world faces many hurdles such as inadequate infrastructure, skills and financial constraints which are all challenges that can be solved and that is why we have proposed a solution through CGI SEEDS program and UNIDO’s SCALEx program. We firmly believe that digital technologies such as AI can contribute to an inclusive digital economy for all. Our idea presents an effective and scalable solution towards realizing the benefits of a digital economy. Our solution focuses on the Sustainable Development Goals 1. No Poverty, 2. Zero Hunger, 6. Clean Water and Sanitation, 9. Industry Innovation and Infrastructure, 13. Climate Action. 14, life below water, and 15, life on land. The solution we’re proposing focuses on using AI to identify pollution within our waters from multiple sources, from agriculture, such as water runoff by the application of heavy fertilizers, or pollution from heavy industries, such as mining or manufacturing. We’re excited that our solution combines different datasets, such as Sentinel-1 and Sentinel-2 datasets from earth observation, ground-truthing data from sensors, datasets from historic sources, as well as other current datasets, such as weather patterns. We combine these multiple sources of datasets, which effectively create a big dataset for us to apply using AI to then predict the likelihood of where pollution will occur at any point in time within our rivers. It will also predict into the future how likely pollutants will affect land use or our water systems. In addition to that, our solution is proposing the use of AI to support precision agriculture and farming. It will also support the identification of renewable sources of energy through scenario modeling and predicting availability of energy sources like solar into the future. Through this forum and through UNIDO, we would love to partner with like-minded people, organizations, and countries that would love to drive innovative solutions for sustainable development, for growth, and for future generations.

Isabel de Sola: Thank you. Wonderful. Thank you, Maddy, or thank you CGI colleagues who are listening online for that example, which is so relevant today. I’d like to hand the baton to Liu Hao from BIT, yes, for a solution from your hometown. Could you tell us where is your hometown?

LIU Hao: I’m coming from China and today I have five minutes to present the AI Green Index. So I will try to finish within five minutes. So when we are talking about the efficiencies that are brought by the AI, we have to realize that AI is a bigger consumer of the energy and the other resources. So if we don’t take any action, AI itself will be a big environment in danger. So well, the regulation for the green AI or the regulation for the AI green performance is not right there. Not only the green performance, even the generative AI regulation. We have AI act in Europe, but for the other regions, we are not quite ready. And the global level, we are still lacking behind. A lot of action are waiting for us. So if we want AI to be green, we need to define the green AI. So our solution understanding that a green AI is a sustainable practice to minimize environmental impact and it continues to improve organizational and social wellbeing. So to make AI green, it is not only AI itself, it’s an ecosystem. It’s hard to consider the infrastructure, the algorithm, the regulation compliance, and also the socioeconomic part. So this is a ecosystem of systematic thinking. We have some tools to measuring, to calculating, but they are not quite enough. So that is why the AI green index is providing a comprehensive index so that we can calculate different parts and we can make the green tool with the guidance and standards in the. AI Green Index, we consider that AI will not only be an enabler, so it’s driven innovation. Yes, it’s a tool, but we will also make AI to help achieve the other sustainable goals. So this is the joint effort and joint product between UNIDO team and Beijing Institute of Technology. So the AI Green Index has five different dimensions, or we call the five pillars. With 18 indicators, okay, with 18 indicators. So you don’t need to read right now because I will share all of the slides. So everything is based on good monitoring and trustworthy data. So it is not only the ranking, you will have good data to show whether it is the performance, whether it is a balance. So we have a variety of users that can use this index. We also have a long list of values that have been created by this index system. I will not read them, and it will definitely help with the digital transformation. The next work will be, we will work on the optimization of future weight. We will also create online tools. So it is not inclusive, it will be inclusive. And we will first, a group of users, a pilot user will work together with us. So the goal is simple. So use the AI Green Index, let the AI to be green, and also serve for the green digital transformation. So if you wanted to know more, if it’s gone on slides, you may contact me.

Isabel de Sola: That was even faster than the five minutes. Thank you so much. And I’d like to turn to Selena from the CEO of Zindi for some thoughts on her youth-led AI innovation in Africa.

Celina Lee: Okay. Thank you. Oh, no. You’re pulling up my slides. Okay. So, my name is Selina. I’m the CEO and co-founder of Zindi. We like to consider ourselves and say that we are the AI innovation and talent factory for the world. And we are a community now of 90,000 data scientists and AI developers from across 180 countries in the world. We run challenges of let me just get to my next slide. We run so when a young person joins Zindi, they’re able to join challenges. So, we host we’ve hosted over 400 challenges. These are addressing everything from the SDGs to business problems. These are real world challenges that come from governments, companies, startup companies. And every time we launch a new challenge, we open it up to our community. Our community competes to build the AI solutions to solve these problems. And every time that they enter one of these competitions, they’re now able to build up their Zindi profile. So, every accomplishment that they make gets added to their profile, which becomes like an online CV or a portfolio of work, which then allows them to, of course, attract the attention of hiring managers and companies that need them. So, I wanted to give just give an example. So, I said we have 90,000, 180 countries. But this is part of just a slice of the community that is in Kenya. We did a study recently of our data on our Kenyan users. And we found that we have about 9,000 people in Kenya. We found that close to 2,000 of them have had some kind of job outcome, some positive job outcome, like a promotion or a new job. And we found that the people who had positive job outcomes, the only thing they had to do, they did not have to win the competition. but if they entered five competitions, they had five use cases that they have worked on and that they joined a team. That means that they’ve formed teams across their peers. They were able to get job outcomes. So I think this was a really exciting finding. And yeah, just in the end to say that, we’re at close to a hundred thousand now. Our goal is to get to a million users. Our vision is to make AI accessible to everyone. So that means everything from the problems that need AI to help them solve them, as well as for young people across the world to have access to those opportunities to build their skills.

Isabel de Sola: Thank you, Selena. That’s really inspiring. And it seems like such a great environment to be interacting with. We have some colleagues here from AWS that I’d like to introduce. I have Sasha Rubel and I’m sorry, Doug Smith as well. Welcome to the conversation and tell us about your AI solutions.

Sasha Rubel: Thank you so much for having us. I will cede the majority of my five minutes to Doug Smith who is acting CEO of Data Friendly Space because it is emblematic of our commitment at AWS not only to democratize access to AI innovation, but to also highlight the stories on the ground that we see as best practices to scale. And I will just say that for us, the big question and the big opportunity of AI and Amandeep was just sharing this earlier over lunch is how do we reduce the space between data, information and insights and life-saving action. And so really excited that Doug Smith was already in Geneva and was willing to share a little bit more about his solution at Data Friendly Space, and Gannett that was the recipient of an award at AWS and with whom we’re very happy to continue to cooperate with.

Doug Smith: Well, thank you for the opportunity. to speak. March 28th, a massive earthquake hit Myanmar, and the devastation was quite significant. Fifty percent of the buildings in Mandalay were either destroyed or really almost beyond repair. It had an $11 billion impact on the economy and probably has meant that we’ll see year-over-year inflation in that region at beyond 34 percent, which is quite high. In the hours as the world began to hear about the earthquake, it was a common story, which is a lack of information, a lack of trustworthy sources, and really general chaos on how it is that the humanitarian sector can respond. We had experienced this over and over and over again since 2018 when we were founded, but we had something else in our toolbox this year. As Sasha said, we launched two years ago a collaboration with AWS, and we call it Gannet, and it had a significant impact on the way that we were able to respond. Within three hours, we had taken the challenges that were on the table, challenges that we had experienced over and over again, and we were able to deploy Gannet. Gannet is a suite of AI-enabled tools that are built for the humanitarian and development response communities. And what it’s allowed us to do within those first three hours is to very quickly mobilize, pull data from trusted sources like UN OCHA, and then generate response reports and analysis that we could then push out to responders. Maybe more importantly, very quickly, once the chaos began to calm, we were able to move that analysis into the hands of local actors, so that they had access to the same information that those responding from Geneva had. That really is one of the powers of AI, to be honest with you. The toolbox has three dominant parts. One is a virtual assistant. It’s a generative system that allows you not just to get results, but to ask questions of the results. And if your question and response hasn’t created clarity, you can follow up with yet another question and have a conversation with the data. That’s the power of AI in the humanitarian space. And fortunately, this is deployed. This is not an idea. It’s not a concept. This is something that we’re actively using on the ground. We heard very quickly from our partners on the ground that they needed actually something that was providing them very fast information within a framework that’s common within the humanitarian space. And so, we put this into something that we call the Situation Hub, and it provides situational analysis. Now, traditionally, we do point-in-time analysis once, twice a year on situations. We were able to update this analysis in the first days of that daily, and to this day, we maintain weekly updates with a human in the loop. So, we take the AI, we build a RAG model on top of a foundational model, and the RAG heavily weights trusted sources coming from UN partners, and then we put a human in the loop to make sure that that automated analysis is curated and accurate. Last week, we actually launched something called Gannett Workspace because we know there are some partners that need private and secure information. And so this will allow people to spin up very quickly their own rag within the foundational model and to query that information. Now, we think that this is really quite a game changer. And the impact, I have to say, is fairly significant, because we’ve already deployed this in Lebanon, now in Myanmar, in Sudan. We are well beyond the MVP. At this stage, what we really want to do is see this scale up to increase the impact. That way, the next time a crisis hits, we can be more prepared to be able to provide actors on the ground with that information. Now let me quickly say that the reviews have been quite good. OCHA used this information. UNDAC and their ANA cell used GANET analysis as their primary tools in their response. And I think that their feedback to us has been very helpful. We have made adjustments. That’s how we work. We’re very agile. And in addition to that, NGOs on the ground also have given us fantastic feedback about ways that we can help them to work with local populations to very quickly get back on their feet, to increase the economic recovery from large-scale events, and to own the response and the data.

Isabel de Sola: Thank you. Thank you so much, Sacha and Doug. We have two more solutions to be presented. I’d like to turn to Jean St-Pierre from Microsoft. Jean-François.

Jean‑Francois Saint‑Pierre: Thank you very much. And thank you for having us. I’m happy today to introduce Microsoft Elevate. And you probably haven’t heard about it because 24 hours ago it did not exist. It was announced at 6 p.m. yesterday. So what is Microsoft Elevate? Microsoft Elevate sees the bringing into one organization of technology support, donations, and sales for schools, community colleges, and nonprofit organizations. It is the successor to and expands upon the longstanding work of Microsoft philanthropies and tech for social impact. I used to be tech for social impact. It’s the team that supports the not-for-profits in the United Nations. More broadly, this is the next chapter for corporate philanthropy and our noncommercial business model. As we have with tech for social impact, we will run this new business with commitments to invest a share of our profits into the nonprofit programs. We announced yesterday that over the next five years, we will donate on a global scale more than $4 billion in donations, AI and cloud technology to schools, communities, and technical colleges, and nonprofits to advance their missions. In addition, is it working? Yes. In addition, Microsoft Elevate will also pursue the next phase of our global skilling programs and initiatives through the Microsoft Elevate Academy. It will help bring AI education and skills to people around the world. In the next two years, Microsoft Elevate Academy will help 20 million people learn an in-demand AI skilling credential ranging from foundational fluency to advanced technical training and working in close coordination with other groups across Microsoft and LinkedIn and GitHub. Microsoft Elevate will deliver AI education and skilling at scale. It will work as an advocate for public policies around the world to advance AI education and training for others. This is in recognition to a changing skills environment. 70% of skills used in most jobs will change by 2030. 30% since last fall. We can see that AI hiring has grown 30% faster than overall hiring. And 78% of leaders are considering hiring for AI-specific roles. And Microsoft is building on existing core offerings. Offerings around curriculum, credentials and training, like AI Skills Navigator, Microsoft Learn, which is open and free for everyone, and LinkedIn Learning, community and recognition, and Data & Insights. And building upon that, we have partnerships, just like the Learning Passport that was built by UNICEF with the support of Microsoft, which has allowed up to 9 million children and young people in 45 countries an access to continuous quality and inclusive education and bridged the digital divide. This is my last slide. Hopefully it comes. But Microsoft Elevate will unlock the opportunity of AI for everyone. Thank you.

Isabel de Sola: Thank you, Jean-Francois. And we have one last solution before we tie this all together. From NTT Data, I’d like to hand over to Manel Martorama.

Manel Martorana: Thank you very much. Thank you very much for inviting us, and let us join this very important conversation on our future. My name is Manel Martorana, and I represent NTT Data, which is the IT services provider from the NTT Group, which is the telco that I think you already know from Japan. We are the sixth IT services company in the world and the third largest data center provider in the world. And also… And also, we are the second digital provider for the United Nations. These are the official figures. I have to say that maybe in the following months we can say we are no more the second, and maybe we are the first. We’ll see next year. We have been working for the United Nations the last 10 years in different agencies and a lot of different projects. So, talking about the challenges that we are seeing in many industries when it comes to is that sometimes the digital divide is becoming even worse or even larger than it was before. For example, let me highlight some challenges. There are a critical shortage on qualified workers in operational environments. Nearly 30% of the world population lives with vision impairments or hearing loss, and 20% have some degrees of discapacity. And despite the improvements in automation, we still see hundreds, millions of workplace accidents over the world. This is a reality. So, the question is, what can we do? But also, what benefits can we get from AI initiatives? And for this reason, we tried to bring into the conversation some examples of real projects and real activities that we are doing in several industries, applying AI, trying to make this digital divide a little bit smaller. For example, what we call the perfect store, that it’s applying AI on the vision of, for example, retail shave and seeing which is the provision that we need to establish for this retail store, or for example, applying in certain industries, large visual models, large visual models where we can see without the intervention of humans, what is happening in a warehouse, for example, but also, for example, for security implication. Also, when there are people in the field that needs to access the knowledge base, for example, for acting in a solar plant or in a distant industry, accessing to the knowledge of the company is not that easy and applying different kind of chatbots that can particularize to the context and to devise the information that is needed, for example, to make any kind of actions on the industry. This makes that every people can access information. And of course, this is one thing that with the application of Gen AI, it’s been a lot easier and a lot of reality than it was before. That’s it. That’s just some examples of the application of Gen AI in different industries that makes really technology more affordable for everyone. Thank you very much.

Isabel de Sola: That’s wonderful. Thank you, Manel. And I, you know, this has been a tour de force around the world and from high and low levels. Let me see if I can summarize some of the solutions that we’ve got. So if we start with the problems, with the challenges that Si Yong articulated for us, there are skills gaps, there’s connectivity gaps, there is a lack of data, a lack of infrastructure, a cloud, unaffordable technologies, the digital literacy piece. So to have an inclusive and beneficial economy, we need to come at it from all of those things at the same time. And I think we managed to do that in 40 minutes. Thank you, everyone, for being so brief. You know, we heard about GAN. So, that’s using data, essentially, to reach vulnerable, marginalized populations. You gave an example of a time of crisis, but the know-how of that could be applicable across many industries, I think, and marketplaces. How do you reach the target population just in time for sales or with bottles of water? We heard about, as well, reaching specific populations like youth to have them learn from each other, learn by doing, experiment with the technology so that they can build the digital skills that they need and literacy, also to create networks amongst them so that they come at problems together. And I think Microsoft’s initiative, congratulations on this announcement yesterday, is squarely in the space of building digital literacy as well. And then, we heard about bringing the AI to the factory floor, so either to enhance the humans. Actually, I couldn’t see the screen, so the visually impaired is now me at 46. But also to help us when we don’t see things, security problems, or to emerge information that could make us more efficient on the factory floor, but at an affordable price. And then, we heard about solutions, the Green AI Index and the SEED project from CGI that are looking at how to deal with some of the unsustainability of our economies, how to deal with some of the perhaps negative impacts of our production of agriculture or of the digital infrastructure and tools that we’re using. So, in a nutshell, Jason, it’s going to take a village to have an inclusive digital economy, I think, or it’s going to take small, large initiatives from all different types of organizations around the world. So, let me hand over to you for maybe a call and a close. Okay.

Jason Slater: Thank you very much, Isabel. And knowing that we’re slightly overrunning and in the spirit of RAPID, I will try to keep this to maybe one minute. minute. So the purpose of this, first of all, also I would just like to echo, thank you so much for those of you who came together and all of those of you who have joined us in this session today. The purpose of this is that today is when we will be making our global call. This is a collaboration between Nido, UNCTAD, UNDCO and Audit, amongst others who are part of the working group. And this is a launch for a global call for solutions aimed at fostering an inclusive digital economy by identifying innovative digital solutions that empower undeserved communities worldwide. This initiative supports the 2030 agenda for sustainable development by promoting equitable access to digital technologies and opportunities. The Global Initiative for Digital Inclusion, this is a call today where we’re looking for solutions that empower marginalized groups including women, youth, small medium enterprises, startups, innovators to build an inclusive, secure and sustainable digital economy. Applications are open to everybody. Innovators, entrepreneurs, NGOs, organizations advancing digital and inclusive empowerment. We’re focusing primarily on five thematic areas. This is around skills, empowerment, enabling policy, innovation acceleration, sustainable digital supply chains to promote digital learning, market access, etc. What will happen to those who are ultimately selected? There will be a judge of us and it will come around from UN and others where we will have selected solutions and our aim is to then promote that during UNGA 80, also going into Nido’s own general conference and others, possibly E-Trade, etc. So that we can help promote and showcase with the hope that we can help you then in deploying some of those within our projects and programs to really demonstrate impact on the ground. So thank you and with that I’m happy to announce that our call is now open and you will have the opportunity to apply in the next few days. Thank you very much Isabel.

Isabel de Sola: Thank you UNIDO for bringing us for bringing us all together and have a wonderful rest of your afternoon and rest of the summit. Thank you.

Agreement points

Digital divide and lack of inclusivity in current digital economy

Speakers

– Isabel de Sola
– Ciyong Zou
– Mattie Yeta

Arguments

The digital economy exists but lacks inclusivity with 80% of companies worldwide lacking web pages and uneven AI uptake across countries and industries

AI is fundamentally reshaping industries, economies, and societies, but the transformation must serve everyone, particularly those historically left behind

Internet usage disparity shows 93% in high-income countries versus 27% in low-income countries, leaving 2.6 billion people without internet

Summary

All speakers acknowledge that while digital transformation is occurring, there are significant gaps in access and participation, with billions of people and most companies still excluded from the digital economy

Topics

Development | Economic | Infrastructure

Skills shortage as critical barrier to inclusive digital economy

Speakers

– Isabel de Sola
– Ciyong Zou
– Jean‑Francois Saint‑Pierre

Arguments

Key hurdles include lack of qualified AI workforce, unreliable data, poor infrastructure, and low purchasing power for AI licenses

Global digital skills shortage threatens to leave 85 million jobs unfilled by 2030, creating deeper digital divides

70% of job skills will change by 2030, with AI hiring growing 30% faster than overall hiring

Summary

There is strong consensus that the lack of qualified workforce and rapidly changing skill requirements represent fundamental barriers to achieving an inclusive digital economy

Topics

Development | Economic

Need for comprehensive ecosystem approach rather than isolated solutions

Speakers

– Amandeep Singh Gill
– Ciyong Zou
– Jason Slater

Arguments

AI solutions work best when emerging from organic digital economy ecosystems, requiring foundational digital public infrastructure, policy frameworks, and strategic investments

UNIDO is building scalable AI platforms and industrial ecosystems through collaboration with 120 partners across 40 countries to make AI affordable and actionable for local industries

The Global Initiative for Digital Inclusion launches a call for solutions focusing on skills empowerment, enabling policy, innovation acceleration, and sustainable digital supply chains

Summary

Speakers agree that successful digital transformation requires building comprehensive ecosystems with multiple stakeholders rather than implementing isolated technological solutions

Topics

Development | Economic | Infrastructure

Similar viewpoints

Both speakers emphasize practical, hands-on learning approaches for AI skills development, with focus on community-based learning and real-world application rather than traditional educational models

Speakers

– Celina Lee
– Jean‑Francois Saint‑Pierre

Arguments

A community of 90,000 data scientists across 180 countries can build AI skills through real-world challenges, with job outcomes achieved by entering five competitions and joining teams

Microsoft Elevate will donate $4 billion over five years and help 20 million people learn AI skills through the Microsoft Elevate Academy

Topics

Development | Economic

Both speakers advocate for AI solutions that democratize access to critical information and capabilities, particularly in humanitarian contexts where speed and accessibility can save lives

Speakers

– Sasha Rubel
– Doug Smith

Arguments

AWS partnership demonstrates how to reduce the gap between data and life-saving action in humanitarian contexts

AI-enabled tools can provide rapid humanitarian response within three hours of disasters, generating analysis reports and mobilizing local actors with the same information available to international responders

Topics

Development

Both speakers focus on environmental sustainability and the dual role of AI as both a solution for environmental challenges and a technology that must itself be made environmentally sustainable

Speakers

– Mattie Yeta
– LIU Hao

Arguments

AI solutions should address precision agriculture, renewable energy identification, and climate action through scenario modeling

The AI Green Index provides a comprehensive measurement system with five pillars and 18 indicators to make AI environmentally sustainable

Topics

Development | Sustainable development

Unexpected consensus

Corporate responsibility in AI democratization

Speakers

– Jean‑Francois Saint‑Pierre
– Sasha Rubel
– Manel Martorana

Arguments

Microsoft Elevate will donate $4 billion over five years and help 20 million people learn AI skills through the Microsoft Elevate Academy

AWS partnership demonstrates how to reduce the gap between data and life-saving action in humanitarian contexts

AI applications in retail, warehouses, and industrial settings can make technology more accessible and affordable for everyone

Explanation

Unexpectedly, all major technology companies represented showed strong alignment on using their resources and platforms to democratize AI access rather than focusing primarily on commercial applications. This suggests a significant shift in corporate strategy toward inclusive development

Topics

Development | Economic

Human-centered AI implementation

Speakers

– Ciyong Zou
– Doug Smith
– Manel Martorana

Arguments

UNIDO is building scalable AI platforms and industrial ecosystems through collaboration with 120 partners across 40 countries to make AI affordable and actionable for local industries

Gannet system includes virtual assistants, situation hubs, and private workspaces that allow conversational interaction with data and regular updates during crises

AI can address workplace challenges including critical shortage of qualified workers, accessibility for people with disabilities, and workplace safety through visual models and knowledge access systems

Explanation

There was unexpected consensus across different sectors (industrial development, humanitarian response, and corporate IT services) on the importance of human-in-the-loop AI systems and designing AI to augment rather than replace human capabilities

Topics

Development | Human rights

Overall assessment

Summary

Strong consensus emerged around the fundamental challenges of digital exclusion, skills gaps, and the need for ecosystem approaches. All speakers agreed on the urgency of making AI accessible and beneficial for underserved populations, with surprising alignment between UN agencies and private sector on collaborative approaches.

Consensus level

High level of consensus with significant implications for coordinated global action. The alignment between public and private sector speakers suggests potential for effective partnerships in implementing inclusive digital economy initiatives. The shared recognition of systemic challenges and ecosystem solutions indicates readiness for comprehensive, multi-stakeholder approaches rather than fragmented efforts.

Different viewpoints

Unexpected differences

Overall assessment

Summary

The session showed remarkable consensus among speakers on the fundamental challenges and goals of creating an inclusive digital economy through AI. The only areas of variation were in implementation approaches rather than fundamental disagreements.

Disagreement level

Very low disagreement level. This was a highly collaborative session where speakers built upon each other’s points rather than challenging them. The lack of significant disagreement suggests strong alignment within the UN system and partner organizations on digital inclusion priorities, but may also indicate limited diversity of perspectives or the structured nature of the presentation format that didn’t encourage debate.

Partial agreements

Similar viewpoints

Both speakers emphasize practical, hands-on learning approaches for AI skills development, with focus on community-based learning and real-world application rather than traditional educational models

Speakers

– Celina Lee
– Jean‑Francois Saint‑Pierre

Arguments

A community of 90,000 data scientists across 180 countries can build AI skills through real-world challenges, with job outcomes achieved by entering five competitions and joining teams

Microsoft Elevate will donate $4 billion over five years and help 20 million people learn AI skills through the Microsoft Elevate Academy

Topics

Development | Economic

Both speakers advocate for AI solutions that democratize access to critical information and capabilities, particularly in humanitarian contexts where speed and accessibility can save lives

Speakers

– Sasha Rubel
– Doug Smith

Arguments

AWS partnership demonstrates how to reduce the gap between data and life-saving action in humanitarian contexts

AI-enabled tools can provide rapid humanitarian response within three hours of disasters, generating analysis reports and mobilizing local actors with the same information available to international responders

Topics

Development

Both speakers focus on environmental sustainability and the dual role of AI as both a solution for environmental challenges and a technology that must itself be made environmentally sustainable

Speakers

– Mattie Yeta
– LIU Hao

Arguments

AI solutions should address precision agriculture, renewable energy identification, and climate action through scenario modeling

The AI Green Index provides a comprehensive measurement system with five pillars and 18 indicators to make AI environmentally sustainable

Topics

Development | Sustainable development

Key takeaways

The digital economy exists but lacks inclusivity, with 80% of companies worldwide lacking web pages and significant disparities in AI adoption across countries and industries

A global digital skills shortage threatens to leave 85 million jobs unfilled by 2030, requiring urgent action to prevent deepening digital divides

The Global Digital Compact provides an action-oriented framework with inclusive digital economy as its central objective, moving beyond principles to concrete implementation

AI solutions are most effective when emerging from comprehensive digital economy ecosystems that include digital public infrastructure, policy frameworks, and strategic investments

Successful AI implementation requires addressing multiple barriers simultaneously: skills gaps, infrastructure deficits, data accessibility, and affordability challenges

Real-world AI applications demonstrate significant impact across sectors including humanitarian response, environmental monitoring, skills development, and industrial applications

Building an inclusive digital economy requires collaboration between public and private sectors, with different approaches for various sectors like health, agriculture, and manufacturing

Resolutions and action items

Launch of the Global Initiative for Digital Inclusion call for solutions, focusing on five thematic areas: skills empowerment, enabling policy, innovation acceleration, and sustainable digital supply chains

UNIDO to continue co-leading Objective 2 of the Global Digital Compact on inclusive digital economy

Microsoft Elevate commitment to donate $4 billion over five years in AI and cloud technology to schools and nonprofits

Microsoft Elevate Academy to help 20 million people learn AI skills over the next two years

Selected solutions from the global call will be promoted during UNGA 80 and UNIDO’s general conference

UNIDO to continue building scalable AI platforms through collaboration with 120+ partners across 40 countries

Development of online tools for the AI Green Index system with pilot users

Unresolved issues

How to effectively measure the digital economy, with current global estimates varying widely between 10-40% of global GDP

Standardization of AI green performance regulations and global-level AI governance frameworks

Scaling successful pilot projects to comprehensive industrial AI ecosystems

Addressing the concentration trend toward large organizations in the digital economy

Ensuring AI solutions remain context-aware and accessible to micro, small, and medium enterprises

Balancing AI’s potential benefits with its environmental impact and resource consumption

Suggested compromises

Adopting a building block approach that prioritizes foundational digital infrastructure before pursuing advanced AI solutions

Implementing human-in-the-loop systems for AI applications to ensure accuracy and local relevance

Focusing on organic ecosystem development rather than attempting to directly compete with major AI players

Combining public and private sector leadership based on sector-specific needs and capabilities

Balancing AI advancement with environmental sustainability through comprehensive green AI measurement systems

Perhaps most alarming is that our global digital skills shortage threatens to leave 85 million jobs unfilled by 2030. Without deliberate targeted action, the digital divide will not just persist, it will deepen, creating a world where the benefits of AI are concentrated among the very few, while the many are left further behind.

Speaker

Ciyong Zou (UNIDO Deputy Director General)

Reason

This comment reframes the digital divide from a static problem to a dynamic, worsening crisis with specific quantifiable consequences. It introduces urgency by highlighting that inaction will lead to exponential inequality rather than maintaining the status quo.

Impact

This stark warning set the tone for the entire discussion, establishing the urgency that permeated all subsequent presentations. It shifted the conversation from theoretical benefits of AI to concrete risks of exclusion, making every solution presented feel like a necessary intervention rather than an optional enhancement.

If you look at global estimates, anywhere between 10 to 20%, some people say that in a short while, this could be up to 40% of the global GDP. But how do we count this? What do we leave out? Is it only ICT spend? It’s something else.

Speaker

Amandeep Singh Gill (UN Special Envoy on Technology)

Reason

This comment reveals a fundamental measurement problem – we’re trying to build an inclusive digital economy without even knowing how to properly measure what constitutes the digital economy. It challenges the assumption that we have clear metrics for success.

Impact

This observation introduced a meta-level complexity to the discussion, suggesting that the challenge isn’t just about creating solutions but about defining what success looks like. It added intellectual depth by questioning the foundational assumptions underlying policy discussions.

If we ask the questions: why are the US and China so successful in deploying AI solutions? It’s because there is a broad base of the digital economy. So those seeds are falling on fertile land… countries need to take a strategic approach to building the digital economy rather than a blind race to match OpenAI, DeepSeek.

Speaker

Amandeep Singh Gill

Reason

This ecosystem thinking challenges the common approach of trying to leapfrog directly to advanced AI without building foundational digital infrastructure. It introduces the concept of ‘fertile ground’ as a prerequisite for AI success.

Impact

This comment fundamentally shifted the discussion from individual AI solutions to systemic thinking about digital ecosystems. It influenced how subsequent speakers framed their solutions – many began emphasizing foundational elements like skills, infrastructure, and community building rather than just technological features.

We found that the people who had positive job outcomes, the only thing they had to do, they did not have to win the competition, but if they entered five competitions, they had five use cases that they have worked on and that they joined a team… They were able to get job outcomes.

Speaker

Celina Lee (CEO of Zindi)

Reason

This insight challenges the traditional merit-based competition model by showing that participation and collaboration matter more than winning. It reveals that skill-building through practice and networking creates economic opportunities regardless of being ‘the best.’

Impact

This finding provided concrete evidence for inclusive approaches to AI skill development, validating the ecosystem approach Gill had outlined. It shifted the conversation toward understanding how to create meaningful participation opportunities rather than just identifying top talent.

How do we reduce the space between data, information and insights and life-saving action.

Speaker

Sasha Rubel (AWS)

Reason

This comment distills the entire AI-for-development challenge into a single, powerful question about the gap between having information and being able to act on it. It reframes AI not as a data processing tool but as a bridge to actionable intervention.

Impact

This framing influenced how Doug Smith presented the Myanmar earthquake response, emphasizing speed and actionability rather than just analytical capability. It shifted the focus from AI as a analytical tool to AI as an enabler of rapid, effective response.

70% of skills used in most jobs will change by 2030. 30% since last fall. We can see that AI hiring has grown 30% faster than overall hiring.

Speaker

Jean-François Saint-Pierre (Microsoft)

Reason

These statistics reveal the unprecedented pace of change in the job market, suggesting that traditional education and training models are fundamentally inadequate for the current rate of transformation. The acceleration since ‘last fall’ shows how rapidly this is evolving.

Impact

These concrete numbers reinforced Zou’s earlier warning about job displacement but added nuance about the speed of change. It validated the urgency of all the skill-building solutions presented and emphasized why traditional approaches to workforce development are insufficient.

Overall assessment

These key comments fundamentally shaped the discussion by establishing three critical frameworks: (1) urgency – the digital divide is actively worsening and will create massive job displacement, (2) systemic thinking – successful AI deployment requires ecosystem development rather than isolated solutions, and (3) actionability – the value of AI lies not in its analytical capabilities but in its ability to enable rapid, effective action. The comments created a progression from problem identification (skills shortage, measurement challenges) to strategic thinking (ecosystem approach) to practical validation (participation over competition, speed to action). This intellectual scaffolding transformed what could have been a series of disconnected solution presentations into a coherent narrative about building inclusive digital economies through foundational infrastructure, community-based learning, and action-oriented AI deployment.

How do we measure and define the digital economy more precisely?

Speaker

Amandeep Singh Gill

Explanation

Current global estimates vary widely (10-20% of GDP, potentially up to 40%), and there’s uncertainty about what should be counted – whether it’s only ICT spend or includes other elements. Better measurement is needed for policy and investment decisions.

How can we ensure AI solutions emerge organically from digital economy ecosystems rather than as isolated implementations?

Speaker

Amandeep Singh Gill

Explanation

Understanding why US and China are successful in AI deployment due to their broad digital economy base is crucial for developing countries to build strategic foundations rather than trying to match specific AI companies directly.

What are the optimal graduation pathways for countries to move to higher levels of digital economy maturity?

Speaker

Amandeep Singh Gill

Explanation

Different maturity levels require different approaches and investments. Research is needed on how countries can systematically progress through these levels.

How can we make AI affordable, adaptable and actionable for micro, small and medium enterprises globally?

Speaker

Ciyong Zou

Explanation

With 80% of companies worldwide lacking web pages and varying AI uptake, there’s a critical need to understand how to scale AI solutions to smaller enterprises that form the backbone of many economies.

What are the most effective ways to address the global digital skills shortage that threatens to leave 85 million jobs unfilled by 2030?

Speaker

Ciyong Zou

Explanation

This represents a massive challenge that could deepen the digital divide if not addressed through targeted, scalable solutions.

How can we optimize the weighting system for the AI Green Index and develop comprehensive online tools?

Speaker

Liu Hao

Explanation

The AI Green Index is still in development and needs refinement of its measurement methodology and user-friendly tools for broader adoption.

What are the best practices for scaling successful local AI solutions to global applications?

Speaker

Multiple speakers (implied from various solution presentations)

Explanation

Several presenters showed successful local implementations but the challenge remains how to adapt and scale these solutions across different contexts and regions.

How can we reduce the gap between data, information, insights and life-saving action in humanitarian contexts?

Speaker

Sasha Rubel

Explanation

This was identified as a key opportunity for AI applications, particularly in crisis response situations where speed and accuracy of information processing can save lives.

What regulatory frameworks are needed for green AI performance measurement globally?

Speaker

Liu Hao

Explanation

Current regulations for AI environmental impact are lacking globally, with only Europe having some AI regulations, leaving other regions without adequate frameworks.