Who Benefits from Augmentation? / DAVOS 2025

Summary

This discussion focused on the benefits and challenges of AI-driven workforce augmentation. Panelists explored how organizations can implement AI technologies while avoiding creating winners and losers among workers. Key points included the importance of upskilling all employees, involving workers in the implementation process, and democratizing access to AI tools.

The conversation highlighted the potential for AI to increase productivity, create new job opportunities, and potentially reduce inequalities if implemented thoughtfully. Panelists emphasized that AI adoption should be worker-led rather than solely management-driven. They also discussed the need for a new social contract to address the changing nature of work in the AI era.

Challenges discussed included the digital divide between developed and developing nations, the need for regulatory frameworks, and ensuring productivity gains translate to wage increases for workers. The panel explored how AI could potentially level the playing field by lowering barriers to entry for high-skilled jobs and enabling workers without formal education to access better opportunities.

The discussion concluded with calls for increased access to AI technologies, human oversight in AI implementation, and the importance of reskilling workers for emerging roles. Panelists also noted the shift from government-led to private sector-led technological innovation and its implications for regulation. Overall, the conversation emphasized the transformative potential of AI in the workplace while stressing the need for inclusive and responsible implementation.

Keypoints

Major discussion points:

– The importance of upskilling and reskilling workers to adapt to AI and technological changes

– Ensuring equitable access to AI tools and training across organizations and globally

– The potential for AI to democratize knowledge and create new job opportunities

– The need for human oversight and control in AI implementation

– Challenges around regulating AI development and use

Overall purpose:

The goal of this discussion was to explore how workforce augmentation through AI can benefit workers and society, while addressing potential risks and inequalities. The panel aimed to provide insights on successfully implementing AI in organizations and economies.

Tone:

The overall tone was optimistic yet pragmatic. Panelists were generally enthusiastic about AI’s potential benefits, but also acknowledged challenges that need to be addressed. The tone remained consistent throughout, balancing excitement about opportunities with awareness of risks and the need for responsible implementation.

Speakers

– Mina Al-Oraibi: Editor-in-Chief of The National, based in Abu Dhabi

– Mohamed Kande: Global Chairman of PwC

– Luc Triangle: General Secretary of the International Trade Union Confederation

– Ravi Kumar S.: CEO of Cognizant

– May Habib: Co-founder and CEO of Writer AI, an LLM enterprise company

Additional speakers:

– Christy Hoffman: Representative from UNI Global Union

AI-Driven Workforce Augmentation: Opportunities and Challenges

This comprehensive discussion explored the potential benefits and challenges of AI-driven workforce augmentation, focusing on how organisations can implement AI technologies whilst avoiding the creation of winners and losers among workers. The panel, featuring experts from various sectors, delved into key aspects of AI adoption, its impact on the workforce, and the necessary steps to ensure equitable and responsible implementation.

Key Themes and Discussions

1. AI Adoption and Workforce Impact

The panellists emphasised the critical importance of upskilling and reskilling workers to adapt to AI and technological changes. Mohamed Kande, Global Chairman of PwC, stressed that access to AI technology is key for workforce augmentation. This sentiment was echoed by Ravi Kumar S., CEO of Cognizant, who highlighted AI’s potential to democratise knowledge and lower entry barriers to jobs.

Luc Triangle, General Secretary of the International Trade Union Confederation, emphasised the desire of workers to be involved in AI implementation processes. This point underscores the importance of inclusive approaches to AI adoption, ensuring that workers are not merely subjects of change but active participants in shaping their future work environments.

An interesting observation came from Ravi Kumar S., who noted that AI adoption is happening more rapidly among consumers than businesses. He pointed out that it has taken only two years for 30% of citizens in many countries to use ChatGPT or similar tools. This disparity suggests a potential gap between personal and professional use of AI technologies, which may have implications for workforce adaptation and organisational change management.

2. Potential Benefits and Risks of AI

The discussion highlighted AI’s potential to increase productivity and create new job opportunities. Ravi Kumar S. observed that productivity has been flat for the last 25 years, with the last productivity bump occurring when personal computers were introduced. He also emphasized AI’s potential to create upward social mobility in jobs if pivoted well.

However, the panel also addressed potential risks. Luc Triangle warned that AI might exacerbate the global digital divide if not implemented inclusively. This concern underscores the need for thoughtful and equitable AI deployment strategies that consider global implications and potential disparities between developed and developing nations.

May Habib, Co-founder and CEO of Writer AI, provided a specific example of AI’s impact on job roles, noting that content writers at Intuit have become content architects due to AI adoption. She also highlighted how HR professionals can use agentic AI to re-engineer time-to-offer processes, demonstrating AI’s potential to transform various business functions.

The importance of human oversight and control in AI implementation was emphasised by Luc Triangle, reflecting concerns about maintaining ethical standards and worker protections in AI-augmented work environments.

Mohamed Kande expanded on the potential benefits, noting that the AI ecosystem can create jobs beyond just the technology sector, including in data centers, real estate, and the energy industry. This insight highlights the broader economic implications of AI adoption and its potential to drive job creation across various industries.

3. Regulation and Governance of AI

The discussion touched upon the complex issue of AI regulation. Ravi Kumar S. suggested that regulating AI output may be more effective than regulating input, presenting a novel approach to balancing innovation with responsible use. The panel also noted the difficulty of regulating AI compared to previous technologies like the internet.

Mohamed Kande noted that the government’s role in AI development differs from previous technologies, with innovation now primarily led by the private sector. This shift presents new challenges for policymakers and regulators in keeping pace with technological advancements.

May Habib suggested that deregulated environments may accelerate AI adoption in companies, highlighting the potential tension between rapid innovation and regulatory oversight.

Luc Triangle emphasized the need for strong social dialogue and workers’ involvement in AI implementation, stressing the importance of updated frameworks to protect workers’ rights and ensure fair distribution of AI’s benefits.

Mohamed Kande stressed the importance of measuring and ensuring access to AI technology, suggesting the need for metrics to assess equitable distribution of AI tools and skills.

4. Change Management in AI Adoption

May Habib emphasized the crucial role of change management in successful AI adoption. She highlighted the need for organizations to effectively communicate the benefits of AI to employees and manage the transition process carefully to ensure widespread acceptance and utilization of AI technologies.

Agreements and Consensus

The panellists largely agreed on the importance of providing access to AI technology to ensure its benefits are widely distributed and to prevent exacerbating existing inequalities. There was also consensus on the necessity of upskilling and reskilling the workforce to adapt to AI-driven changes in the job market.

Both Mohamed Kande and Ravi Kumar S. viewed AI as a tool for increasing productivity and creating new job opportunities, emphasising the need for workforce adaptation through upskilling and reskilling.

Luc Triangle and May Habib highlighted the importance of involving workers in AI implementation and recognising their potential to contribute to and benefit from AI technologies.

Differences and Unresolved Issues

Despite overall consensus, there were some differences in approach, particularly regarding AI regulation. While Ravi Kumar S. suggested regulating AI output for responsible use whilst allowing innovation, May Habib implied that deregulation might accelerate AI adoption in companies.

The role of government in AI development also revealed subtle differences, with Kande implying a more active government role in ensuring access, while Kumar suggested a more hands-off approach focused on output regulation.

Unresolved issues included how to effectively regulate AI given its rapid development and private sector-led innovation, how to bridge the global digital divide in AI adoption and benefits, the timeline for widespread AI deployment in mid-sized companies, and how to ensure productivity gains from AI translate to wage increases for workers.

Conclusion and Future Directions

The discussion concluded with calls for increased access to AI technologies, human oversight in AI implementation, and the importance of reskilling workers for emerging roles. The panellists emphasized the transformative potential of AI in the workplace whilst stressing the need for inclusive and responsible implementation.

Key recommendations included prioritizing AI literacy training and tools for all employees, involving workers in AI implementation processes, ensuring widespread access to AI technology across different regions and socioeconomic groups, and focusing on effective change management strategies for AI adoption.

The conversation highlighted the complexity of managing AI’s integration into the workforce and economy, suggesting that a multifaceted approach involving various stakeholders may be necessary for effective AI governance and implementation. As AI continues to evolve and reshape the world of work, ongoing dialogue and research will be crucial to address emerging challenges and harness the technology’s potential for the benefit of all.

Mina Al-Oraibi: Good morning, thank you for joining us for this session. My name is Mina Al-Araibi, I’m the editor-in-chief of The National based in Abu Dhabi, and I’m delighted to be leading our conversation today, who benefits from augmentation? Clearly, we are at a moment where AI in all its uses, but particularly generative AI is on everyone’s minds and what it will mean for all industries, job creation, job displacement, job transformation. And what we’re going to do here for the next 45 minutes is really talk about how that rapid development of technology, tools, cost assumption and so forth will affect workforces, but also how we can make sure that not only it won’t create further inequalities, but hopefully reduce some of existing inequalities. This discussion is part of the WEF Jobs Initiative that all of us are part of in one way or the other, and it’s a great initiative really thinking about the future of jobs. If you haven’t seen it yet, highly recommend reading the future of jobs report 2025 that just came out this month, with a thousand leading employers talking about what the changes in jobs and that disruption. Now, of course, we think about AI and we think about technology, but really it’s part of a whole host of other developments, increasing cost of living, demographic shifts, geoeconomic fragmentation and so forth. And in that report, it highlights the fact that between 2025 and 2030, so in the next five years, 22% of today’s total jobs will be impacted by some sort of structural labor market changes. So you can follow this conversation online for those of you who are online using hashtag WEF25. To help us in this conversation, we have a great panel. I start at the furthest end of my left, Luke Triangle, General Secretary of the International Trained Union Confederation, Ravi Kumar, CEO of Cognizant, May Habib, co-founder and CEO of Writer AI, which is an LLM enterprise company, and Mohamed Kanda, an organization, of course, well-known to many, global chairman of PwC. Okay, so Hamad, I wanna start with you. As organizations deploy workforce augmentation and adopt more and more of AI technologies, how can they avoid dividing their workers between winners and losers? I might mention three things.

Mohamed Kande: The first one is really around upskilling. And the upskilling needs to be as broad as possible, meaning that all employees are engaged in the upskilling journey within an organization. And when we say upskilling, it’s not just about what is the technology, what is Gen AI, what are the tools? There are two other questions that need to be answered. The second one is also, when do they need to use it? It’s very important not only to understand what it does, but when to use it, but also how to use it. So we are very comprehensive when we think about the upskilling journey of the employees to make sure that they all understand what the benefits of the technology. The second aspect is really around making sure that we listen to them. And the reason why it’s important to listen to the employees because the benefit of augmentation, the people that know how to better get value out of augmentation are the people doing the job. So it’s a good thing to ask them how they think they can get the benefits of augmentation. It shouldn’t be coming from management, it should be coming from the employees. And we call it citizen-led development because the ideas about how to benefit from augmentation is coming from the employees, not from the leadership. The third one is around exposure. You know, people fear what they don’t understand. So exposing them to the technology, putting it in their hands makes a big difference because now they understand that they can use the technology to augment themselves and to decide how to get the augmentation. They’re not being told about the augmentation themselves. And you know, I mentioned earlier last year when I was here at the WEF that the concept of creating a digital colleague that they work with, when you think about having a digital colleague, you don’t fear your colleagues, you partner with them. So we are actually asking people to partner with the technology. I want to pick up on this point of actually getting the workers who are working on a particular issue to problem solve themselves and think about adoption. But that requires also a certain level of understanding. So is that training and how do you implement that if you’ve got an organization or even a government entity? That is why the first step is self-skilling because we have to expose them to what it is, as Adam mentioned, and making sure that they know what the technology does, when to use it, how to use it, and then give them the benefit of providing inputs about how to drive the benefits of augmentation. There is this fallacy of believing that the benefits of augmentation have to come from management. No, it has to come from the people. And then it’s them adopting it. That’s how you accelerate adoption.

Mina Al-Oraibi: Mm-hmm. May, I want to turn to you because there’s been a lot of consideration about this that, you know, where does it come from? Is it manager-led? Is it, again, you know, employee-led? But there’s also governments. Governments are really concerned about this, especially as we see it as a time of geopolitical fragmentation and also changed economies. So last September, G7 labor ministers came together and issued a call for an AI workforce agendas for different countries. Should it be government-led? How much involvement would companies want government to be involved in?

May Habib: I think the reality is the most innovative companies in the world are the ones who are leading the drive to upskill their employees and introduce AI. So I think the train has left the station that this is a private sector-led transformation, but people are hungry for these AI skills. We are a LLM company. We build technology, but we are definitely in the business of change management as well because if you don’t play AI like a team sport, you just don’t get the transformation. So I definitely agree with everything Mohammed said. From a G7 perspective, there are two things that came out of that blueprint that definitely apply to us and I think tech companies should take very seriously. One is this recommendation that AI, and whether it’s agentic or AI applications, be really driven by transparency and by human oversight. And that is both a technology challenge and a product challenge for LLM companies like ours. On the technology side, providing transparency for us has meant really innovating in terms of how you allow solutions built on LLMs to be more deterministic in how they kind of show proof of work to that person who is doing the inspection. So we’ve pioneered a category called graph-based drag that has allowed us to really provide that level of detail, that level of citation. So very important outcome of the action plan. And the second big thing that applies to tech companies that came out of the G7 is this real impetus to involve all levels of an organization in the AI transformation. And for us, that has resulted in just very tactical, low-tech things as well as the high-tech. So office hours and AI days and all hands and trophies and awards and just really gamifying this muscle memory change. It might be a fraction of the time to do it the AI way now or to let the AI agent take care of it, but it actually is a transition. And I do think employers can understate the importance of change management here. It’s like all of us needing to relearn how to read and write. It is so fundamentally different, but when you’ve done it right, our top employees are 10x the productivity. So that is really the kind of outcome that we don’t want to exclude people from. But I mean, what you say is wonderful, but for many companies who are not so well-versed in the technology itself, it’s almost at this point, they know it’s here, there’s no going back, so forth, but there’s also a sense of if they’re excluded,

Mina Al-Oraibi: if they don’t understand the concepts. I mean, to Mohammed’s point, the upskilling, but also understanding the concept and access. So if it’s a company going through a transformation and different industries are represented here, I mean, where do you start thinking about that change management?

May Habib: Yeah, so for us, we like to find the change champions and really put them on a pedestal. We work very closely with Uber and we introduced AI agents to a customer support force of 40,000 people. And there was a guy there, Michael, who is a non-engineer, who really bridged being able to talk to product and engineering as well as the support folks, and did things like Feed Your Mind Fridays, and just gamified it, made it fun, took the fear out of it, as Mohammed said. And so, not every organization is gonna have 100 Michaels, but we’ll help you find five, right? And really helping them pave the way of what it looks like to get others to really embrace the technology. Ravi, I wanna ask you, because a lot of, at least in our industry, for example, in the media,

Mina Al-Oraibi: the way we’re thinking about AI adoption, and have already started to get AI into our newsroom, is thinking about very rudimentary basic skills. Some of those skills are what we would have given our either interns or graduates who have just recently joined or even those who don’t have a formal education. And they would learn the industry through those particular jobs. And those basically are gone now. So how do you make sure that those who are just at the entry level are learning industries and their core basics? Or is that an old way of thinking? Is that no longer necessary?

Ravi Kumar S.: Yeah, so thank you for that. I mean, terrific comments from both of you. You know, I wanna step back a little and talk about why work is organized the way it is. This is evolution from the industrial revolution. And workplaces and workforce, even with digitization of work over the last 25, 30 years, has remained pretty much similar. You know, there was a scientist in the 50s, Alan Turing, who spoke about how machines should be as powerful as humans. And that was a trap. It was a trap because it never improved, it never gave an opportunity for productivity improvement for humans. I mean, you replace humans for work, and most tech disruptions which have happened so far have attempted to replace humans versus augmenting humans or amplifying the potential of humans. This is a unique technology. Going back to what you said, it’s going to disrupt a knowledge worker, it’s going to disrupt workers who do repetitive tasks. And we’ve never had a situation of that kind. So it’s a genuine opportunity for us to disrupt the work template, the template which came from the industrial revolution. The industrial revolution template was, you educate yourself for the first 25 years, you work for the next 50 years, and then you retire. That’s the kind of thing that we’re trying to do. And then you retire. That was the template. That’s actually in some ways up for disruption. So what does this technology do which makes it so disruptive for work workplaces and workforce? It’s a technology which will diffuse very fast. It doesn’t need the skills to access it. The interface is natural language. If the diffusion is fast, anybody can access it. So the entry level to new jobs is going to be lower. The gap within an occupation, the gap between occupations is going to shrink. Interestingly, we did a study at Cardizen, and we have knowledge workers. We have 350,000 of them. The people in the lower percentile actually benefited more than the people at the higher percentile. So at the bottom 50 percentile, we actually had 35% productivity improvement using AI tooling. At the upper percentile, we actually had only 15%. So it’s a leveler of sorts. It’s going to level the high performers to the low performers. It’s kind of a leveler. So if you put all this together, you get a sense that you can create upward social mobility in jobs if you pivot it well enough, not to replace humans, but to augment humans. So we did a study at Cardizen called New Work, New World. We took 18,000 tasks. We took 1,100 occupations. We looked at the exposure score of these tasks to AI, and AI is evolving. I mean, it’s a general purpose technology. It’ll be better in the future than what it is today. It’ll be pervasive. So we know that you should look at it from a template that what it does today to what it does in the next five years. And we created these exposure scores on AI, and we also created friction scores. Friction scores are what would be the new tasks humans would do if you tool them with AI, and what does it take to reskill them? If the friction score is high, it takes you much longer. If the friction score is low, it takes you less. If the exposure score is high, it disrupts your job. So if you pivot in that direction, I think you can really create upward social mobility. You don’t need formal education to get into jobs. You can intertwine, I mean, the capability set needed today is at the intersection of computing data skills and the functional job. If you push that into the K-12 ecosystem, you can actually create individuals who could power their functional jobs with AI tooling, and therefore augment it and not actually replace them. I mean, that’s very important. I mean, I live in the United States. In spite of the United States being a top-developed nation, four million people leave jobs every month. Four million people leave jobs every month, which is one-third of the U.S. workforce in a year. Equally, there are seven million open jobs every month, and these are two different swim lanes, and there’s no bridge in between, and that’s because the bridge to these high-paying jobs is formal education. And the people who are on these jobs, the people who are leaving these four million jobs are in hospitality, hotels, airlines, healthcare, and they don’t have access to those jobs. They are stuck, rolling from one to the other. So we have a unique opportunity to build that bridge, create that upward social mobility using this tooling. If we pivot this well, don’t fall into the Alan Turing trap, as I spoke about, and create shared prosperity. I mean, we have a shot at it. We have a shot to break the template, the Industrial Revolution template, which really created the bridge, which really created the divide, if I may, and we can bridge it. So I’m very, very excited about the prospects of where we could get to.

Mohamed Kande: Can I add one thing? Yeah, sure. You know, Ravi, what you just said, because this can happen within nations, but across nations, I truly believe that this, when you think about the use of AI, what you just described, how do you democratize knowledge? Absolutely, absolutely. And this is, as you said, as you absolutely bang on,

Ravi Kumar S.: you are lending expertise on your fingertips. You know, the past revolutions lent information on your fingertips. There’s knowledge. This, you’re lending knowledge on your fingertips. Therefore, the entry barriers are gonna be broken. And think about the digital divide between what we call today the global South,

Mina Al-Oraibi: which I’m hoping that one day is going to go away, because there’s no such thing. Yeah, absolutely. And you don’t need a four-year degree to access high-paying jobs. But you do at this point. So I think in the future and the near future, we can see where a lot of these changes start to take place. But there is the first mover advantage, and there is the fact that today, and in the past at least five years, whoever’s had access to the data sets, the technology, I mean, it’s resource, be it financial data or otherwise, or actually the skilled workers or the champions that May spoke about that can be skilled. And they become the triggers for a lot of this change. But that is also creating a divide between those who are still trying to figure it out and don’t have access, don’t have the resources. It’s still quite expensive to do a lot of this. So, you know, I think you raised another important point. There are some jobs where the trigger to change is going to come because the consumer is changing.

Ravi Kumar S.: I mean, we We had another research report where 55% of consumers in the age group between 18 to 44 are going to access AI tools to either learn or to buy or to use new things. Now if you are on the front line of a retail company or front line of a business and you are addressing those customers, consumers, you need to get up to speed with what they are trying to do and therefore you have to equip yourself with skills which will support this new consumer. I mean AI adoption in consumers is much, much faster than AI adoption in businesses today. So I think there’s going to be a need, going back to what you said, which is to reskill the workforce irrespective because the consumer on the other side is going to change. So I think all jobs, it’s just not knowledge workers. Jobs which are not even knowledge workers, the people who are doing regular routine jobs, they have to be equipped if you want to create that mobility and create that upward mobility in jobs.

Mina Al-Oraibi: Now Luke, that’s leading to a lot of changes. I mean what we’re hearing from Ravi, this idea of the complete disruption of how we work based on industrial revolution, you’ve spoken about this, you’ve written a lot, you’ve called even for a new social contract because a lot of how we organize society is based on norms that are changing. So one, is it realistic to see that there could be a new social contract? But two, how do we make sure that all these benefits that we’re talking about go to the workers and benefit them rather than it just really being about managers and shareholders leading this and benefiting largely financially?

Luc Triangle: Well, if we want to make this whole evolution into a success story for humanity and for the society, we will have to include everyone from the design and not only when it’s going to be implemented. So in that sense, workers’ involvement, like Mohamed said, and workers’ representatives’ involvement is key in the success of implementation. That’s my first point. Secondly, I think we believe, as global labor movement, we organize workers and trade unions in nearly 170 countries, so for us this is one of the hot topics currently in our debates how we have to deal with it. It’s clear that workers worldwide are not scared of new technology. They are scared if their company remains working with old technology. It’s the way how you implement it. So we want to be on board in all these debates. Workers want to be part of the process in their companies, in their sectors, in their countries, because there are risks. Let’s face the reality as well. There are jobs disappearing currently already, but there is also an augmentation in the jobs. The speed of work is increasing, the number of tasks that people have to perform are increasing all because of technology and AI. So in that sense, what we call for is indeed strong social dialogue, strong workers’ movement in the implementation of all these processes, because we have to deal with some questions. Algorithmic management, how are we dealing with that in our company? To what extent do we use it? I like the expression of human in control. I think that should be a leading principle. I’m scared and we are scared when the future of AI will be dominated or written by AI itself. So that’s going to be the challenge, that’s the box of Pandora that one day will be opened and how are we then still in control as humans? So we need to take that into consideration, how we can avoid that evolution of AI. I’m not an engineer, but I see the risks here for jobs, but also for companies, for the society as a whole if we go into that direction. So there needs to be some boundaries on AI implementation. I think that’s a crucial element and we see that already in a number of countries and also the G7. I was involved in Italy in the G7 declaration, also in the G20. We have seen that in Europe with the AI Act. We see that also in a number of other countries, UK and others, where they also now take initiatives on boundaries of AI. So all this creates uncertainty today and indeed, you have to see this in a global context. It’s not only about AI, it’s not only about digitalisation, we also have the climate agenda, we have the growing inequality in the world, we have the geopolitical conflicts. So our people, the eight billion and more today on earth, they want to have certainty, they want to have a new perspective and in that sense we indeed are currently promoting the idea and the principle of a new social contract for the world. We are moving to the World Social Summit later this year in Doha in November, where we believe that the world needs to have a follow-up after the social development goals, the SDGs, beyond 2030 that focus now on a contract that gives security on jobs, on wages, on living wages, on rights, on social protection, on equality, on inclusion. And the whole evolution of digitalisation fits into that security that we need to give, that certainty. We need to not scare people, but we need to take them on board and give them a guarantee that their voice and their concerns, also on skills, will be taken up. On skills specifically, we say in our demands on the new social contract, let’s anticipate. That means we know that the jobs of tomorrow, or the day after tomorrow, will be different with the jobs that we have today. The question is what do we do today to prepare our workers for the jobs of tomorrow? The scariest environment for a worker is that he knows that his evolution is taking place and it’s not taking up. And skills indeed has to go beyond the manager’s level. Everybody needs to be taken on board, that’s also what you said. So we need to take the whole society on board in our upskilling efforts and we need to do much more than what we do today to make sure that everybody’s on board in this revolution.

Mina Al-Oraibi: Your point about the fear of exclusion being paramount to this, but please Mohamed.

Mohamed Kande: I want to add one more point because when we think about upskilling, you have the upskilling and how people use AI to do their job better, to augment themselves, etc. But you also have the reskilling of the people that are losing the job that they’re doing too. Reskilling. Because here the thing, the numbers that Ravi was giving, that every month you have people departing the workforce, but every month you have new jobs that are being created. What we have seen in the last industrial revolution is that technology actually net-net created new jobs. I mean the last one, you remember the mobile internet when it came out? There’s one job that did not exist before the internet, that was created in every single nation in the world, the website developer. Millions of them. Now, you have jobs that are being created in your company yourself. Why? Because you’re building large language models. So the question is that when you look at what are the jobs that are going to be impacted by AI, but also what are the ones that are going to be created to support even the development of AI, or the implementation of AI, the adoption of AI, and to understand how do you take employees, not only to upskill them to use AI, but to reskill them to do the new jobs that are going to be created. There will be new jobs that are being created today, I’m sure that you have the same thing. We have also more than 300,000 people around the world. We have job openings that we cannot fill. Across the world today, we cannot find the workers, and we are going to use the reskilling of the workers for us to be able to even augment our workforce, not augment when it comes to use of technology, but augment our workforce because we need more people. And all this driven by technology.

Ravi Kumar S.: I think there’s one other aspect which is productivity-led. If you really look at it counter-intuitively, productivity has been flat for the last 25 years. The last time there was a productivity bump was when the personal computers came into picture. So augmenting humans with more tooling so that they’re more productive, they get more wages. As they get more wages, they create upward mobility. And that in turn actually leads to growth without, I mean it’s deflationary, growth without an inflationary thing. I mean today, that’s the biggest issue. As jobs increase, inflation increases because the wages are not increasing at the same level. I mean, that was my point. That’s what I wanted to pick up on what you said, because actually increasing productivity

Mina Al-Oraibi: is not always increasing in wage rises. No, not at all. There’s more being demanded from workers at the same wage. So how do we correct that, I mean, by the organization? If you see, these are not the figures of the trade unions, but these are figures from ILO and from other international research. The labour income share in the growth of GDP is declining.

Luc Triangle: So that means that the growth in GDP, the growth in wealth, the growth in productivity is not shared with those that create the wealth and the increase in productivity. And that’s also an issue that will create, at a certain moment in time, today already, tension and social unrest. And so we need to make sure that everybody’s on board in all ways. And that means also sharing the benefits of this technology, also financially. Let’s be honest and let’s be clear on that point. So that needs to be also recognized. I think it’s the foundation to create a pivot on increase in wages. You know, if you have a productivity increase, and as you rightly said, the value could transfer to the one who’s actually making the tools, to the one who’s using the tools. I mean, we have to pivot on the direction of using the tools. But at least you have an opportunity to improve wages, because you’re going to see uplift of productivity.

Mina Al-Oraibi: May, how do you see it?

May Habib: The practical realities is we are democratizing who can become a builder. And we do value technical skills pretty highly in knowledge work. So we’re definitely seeing folks without technical or engineering backgrounds being able to engineer pretty sophisticated, agentic types of flows. There are a number of customers where folks in HR were able to re-engineer time to offer with agentic AI. So being able to more quickly get contracts out the door, more quickly get people into their seats. And they’re able to build these applications without actually having engineering support. At Intuit, content writers are now content architects. It’s a large technology company we work with, where because they are using agentic AI to orchestrate so much of the work they were doing, they have more leverage in their organizations. The work that they’re doing is more strategic and is more widely applicable as well beyond kind of the specific domain they were in. So I’m very optimistic. I think we are going to see folks who embrace the AI tooling and get skilled, get that commensurate and well-deserved kind of increase relative to the leverage.

Mina Al-Oraibi: Okay, final thought, because I do want to give a chance for questions.

Ravi Kumar S.: I just want to add a very quick thing. I mean, again, if you go back to the Industrial Revolution to now, solving problems was the endeavor of work. So therefore, the STEM skills actually got more prominence than other skills. This is a leveler again, because if problem solving is going to be done by machines, or jointly done by machines and humans, the next human endeavor is going to be about finding the next big problem. And therefore, you’re going to find HR professionals, anthropologists, sociologists, you know, getting access to high-paying jobs, because that’s going to be the next human endeavor. Okay, I want to open the floor for questions. If you have a question, please indicate and we’ll get a microphone to you.

Mina Al-Oraibi: Okay, good. Hammad, I’ll let you continue your thought. You were going to pick up what Ravi was commenting on, you know, this idea of the jobs and the high-skilling jobs, because that opportunity is going to be open. I mean, again, in this transitional phase, and we are really in the transitional phase in different industries, how do you ensure that the access is there and the time is created? Because again, we put so much emphasis on productivity and creating that time to up-skill, to get people to brainstorm, how can this job be done completely differently? How do we ensure that?

Mohamed Kande: Your point around access is critical, right, because we had the same thing with the beginning of the Internet, the whole concept of digital inclusion. Because you know that at one point before the launch of the third-generation wireless networks, most of the people that got access to the Internet were actually using computers. We’re talking about the personal computers, that’s it, right? Not everybody in the world, and it’s not even a difference of the global South or not, it was a matter of income, ability to have a computer, yes or no. And now, in the early 2000s, here comes the third-generation wireless networks, and now the cell phone becomes the primary device to access the Internet. That is a time when we democratised access to information, not knowledge, but information. And the cost of accessing information dropped completely. But it took some time, to be very candid, because the Internet was, when we all had access to it, was in 1993, 1994, around that time, and that happened in 2001. I’m hoping that it’s not going to take eight years. It took seven years for the Internet to get to 50% of what it was. Exactly.

Ravi Kumar S.: From 1994 to 2001. Yes. We are just hoping that it’s not going to take seven years. It’s actually taken two years in many countries to have 30% of citizens using ChatGPT or tools

Mina Al-Oraibi: of that kind. Yes, adoption rate has been incredible. Adoption rate has just been incredible. Which, again, highlights that people are not necessarily scared, they’re interested, but they just want to make sure that they’re included. In fact, we have a reverse problem. Businesses are taking longer to catch up. Yes.

Mohamed Kande: Than people. It’s good to have this whole around user adoption. The adoption of Genia is going to be citizen-led, not business-led. That’s what we’re finding. Luc, you wanted to come in.

Luc Triangle: Well, the point, and I think it was you, Ravi, who mentioned it, the global digital divide, I think we need to discuss that as well. How do we take the world on board in this evolution? Because at the moment, I’m concerned that this evolution will be an evolution for the happy few already at global level, happy few including countries, global norths in majority. When we talk about access to AI, let’s make the connection to access to Internet. In a continent like Africa, where the majority of the people don’t even have access to electricity, what are we talking about? So I think, and that’s also the new social contract, we need to take the world on board in this evolution. Otherwise, this is another engine for growing inequality and growing instability and growing geopolitical conflicts and tensions. So I challenge major companies like yours and others, please take that dimension also on board on where do you invest, where do you look at, and how do you take on board those regions and those countries for which Internet is still a challenge to access today. And honestly, looking forward in this century, there are so many challenges by the end of this century, but we will not have a stable end of this century if that dimension of global inclusion is not taken on board, also not in this AI debate. And honestly, I think the speed of AI evolution and the dimension and the impact of AI is in no way comparable to the low impacts, if I may use that word, of Internet. This is going to be a multiplication of impacts in comparison with what we have seen in the 80s with Internet. So I think we have all, and certainly global companies and governments, have an important role to play to make sure that the global context is taken on board in what we discuss today.

Mohamed Kande: Mohamed, if you can respond to that, and then I have a question for the audience. So, you know, look, I totally agree with you. There is one thing about AI, when we talk about AI, there is the use of AI. Everybody now, do you have it in your hands, can you use it, can you adopt it? But when you think about the AI ecosystem, AI is very different from what we had with the Internet. The Internet was digital, the world was digital. AI is digital and physical at the same time, that’s a big difference. So when we think about job creation in that AI ecosystem, you create jobs in developing large language models, you create jobs in data centers, you create jobs in real estate, you create jobs in the energy industry. You have more innovation and renewable. So in that ecosystem of AI, jobs are going to be created, more jobs than we have today. Because of the domain, it’s not the use of AI. It’s the enablement, the building of AI. It’s not just large language models. It’s a physical infrastructure that we need. That’s the first and the second thing. That physical infrastructure doesn’t need to be in the West. Doesn’t need to be where the people are building the large language models. Now all countries in the world gets to participate. Countries that have the ability to provide energy, to build data centers, whether they have people that are using AI or not, can create jobs. So everybody gets to participate around the world. And that level of participation will be both in the digital world and in the physical world. That is why for us, we think that only good things will happen because of that. Because Africa gets to participate. Southeast Asia gets to participate. Why, because they can build data centers. So the West can actually build large language models. So it’s a very, we look at AI not just a technology, but the ecosystem that it’s creating because it’s now all being connected. So there’s a lot of hope for that. Please, you had a question.

Audience: Go ahead. Hi, I’m Christy Hoffman from Uni Global Union and we represent workers in service sectors around the world. And I just, you know, putting your point about the data centers aside, because I think that’s also a bit of unequal benefit from AI if you’re talking about the workers who can build the data centers versus getting that enhanced productivity at work. But I wanna go back to this adoption question because you did say how fast people are using AI. We’ve talked to our members around the world. And you do see two tracks, like very little gen AI used at work in the Global South. Sort of like, what’s the big deal? Why are we worried about this? It’s not really happening where we live. But equally, even North America, Europe.

Mina Al-Oraibi: Sorry, your mic, I think they, yeah. Oh, sorry.

Audience: People are using AI individually. But a big issue is that without guidance or rules or, you know, direction from their companies, but the level of AI deployment by companies is still pretty low. It seems, you know, that’s my impression from having read about it. And also our members would, they say, you know, we are using it on our own. You know, maybe some use ChatGPT and some use Cloud and some use Perplexity or whatever. But we’re not getting that, you know, it’s not a company-wide deployment in any kind of way, except in the very biggest companies. So I’m just wondering, I know there’s predictions. What’s your prediction in terms of how, when, like, mid-size companies are gonna start to deploy AI with all the expense that that, you know, involves? Because it is an investment. And seeing the, you know, I think this is where some of our concerns are. Certainly those mid-level administrative jobs, which do have a lot of tasks. We represent bank workers. A lot of their work can be replaced by AI. There’s a lot of anxiety among many jobs where there is not only augmentation, but there is automation that’s on the horizon. So what is your foreseeable future about when that adoption will really pick up? I know many people are predicting 2025, 2026, but I’m just interested in your prognosis. Rafiq? Can I, I’ll probably start and you can add.

Ravi Kumar S.: I think regulating AI is gonna be one of the most difficult things to do. I mean, the internet got regulated because the internet was, in many countries, controlled by regulators what went in. By the government, yeah. By the governments. And then you had a way to take it and disperse it. The builders can build it without checking with the government, whatever. I mean, there’s no, you know, the builders can build the cycles of innovation without getting to the regulators. So effectively, the process of building is gonna be much more superior in this way. So I kind of acknowledge and worry about what you’re saying. And therefore, it’s important for regulators to look at it, to enable it, and to control it. I mean, this is not just about controlling, this is about enabling and controlling it. Both are gonna be very important from a regulation perspective. The second is, I think we’ve all looked at this from a context of, should you regulate the models or should you regulate the output coming out of the models? Should you regulate the input coming, should you regulate the input into the models? I think the end state will be regulating the output versus regulating the input. It will allow you to innovate, but it’ll allow you to use it in a more responsible, meaningful way when you do it on the other end. I’m going to take advantage of your pause there because we’ve only got three minutes left. So I wanted to ask you all one question,

Mina Al-Oraibi: which is what’s one concrete measure that you’d advocate for that you’d want people listening in and present here to really push for in order to make sure that augmentation is as beneficial to everybody as possible? So, Hamad, I’ll start with you.

Mohamed Kande: For me, it’s access, access to it. That’s because you won’t have augmentation if people don’t have access to the technology. And access to the technology can be measured because what you have to bring to the hands of people.

Mina Al-Oraibi: May. What percentage of your company has built an app or an agent themselves personally? I would say access and re-skilling. Which is a combination of May and…

May Habib: Well, I’m very specific. It’s like, you’ve gotten AI literacy training, then you’ve gotten the, here are the tooling 101, then did you apply it to your work? Because that is when you really understand the power of it and that you’re in charge, right? And that folks of all sorts of abilities, if they’ve got context and business understanding, are able to build something that really benefits and improves productivity. Luke, what would be your… For us, not only human oversight,

Luc Triangle: but actually human in control. Because then you can steer the process to a certain extent and deal with the negative consequences or the negative impact that might occur through the implementation of this Gen AI. So that’s for us, the most important topic.

Mina Al-Oraibi: And just, I mean, because you guys were all really prompt, so it gives me another question. Just to pick up on this point of boundaries and who determines the regulation of the output, which I think Ravi laid out quite well. I mean, Mohammed, do you see the possibility of regulating the output? And is it going to go back to governments? This is going to be very interesting because in the past, some of the very fundamental technologies

Mohamed Kande: were actually developed by the government. When you think about it, the internet was created and developed by the government and they created from the onset. They could regulate it and then they made it available to the private sector, 1994. But it was government led. Today, the AI innovation is not government led, it’s private sector led. That is what, there’s a big difference between what happened in the last industrial revolution, the one that you have today. And now with the question, where is the government going to play? And May, where do you see that regulation?

Mina Al-Oraibi: You’re really at the helm of a lot of this stuff. So how do you see it?

May Habib: We have an office in London. We’ve got multiple offices in the US. We’re definitely seeing American companies, I think because it is a deregulated atmosphere and a very pro AI atmosphere, just have a lot more interest, acceleration into bringing AI into their companies. I’d like to see Europe catch up.

Mina Al-Oraibi: Okay, so thank you for an incredibly interesting conversation. Access has been fundamental in how we think about augmentation, involvement of workers, involvement of your organization across the board. It’s a team sport. Very important to think about that. But also this idea of human in control and where does the human excel in this? Please thank our excellent panel and continue to follow the jobs initiative. Thank you. Thank you. Thank you. Thank you. Thank you.

Agreement Points

Importance of access to AI technology

speakers

– Mohamed Kande
– Ravi Kumar S.
– Luc Triangle

arguments

Access to AI technology is key for workforce augmentation

AI can democratize knowledge and lower entry barriers to jobs

AI may exacerbate global digital divide if not implemented inclusively

summary

All speakers emphasized the importance of providing access to AI technology to ensure its benefits are widely distributed and to prevent exacerbating existing inequalities.

Need for upskilling and reskilling

speakers

– Mohamed Kande
– Ravi Kumar S.
– Luc Triangle

arguments

Upskilling and reskilling are crucial for workforce adaptation

AI can democratize knowledge and lower entry barriers to jobs

New social contract needed to address AI’s impact on work

summary

The speakers agreed on the necessity of upskilling and reskilling the workforce to adapt to AI-driven changes in the job market and to ensure workers can benefit from new opportunities.

Similar Viewpoints

Both speakers view AI as a tool for increasing productivity and creating new job opportunities, emphasizing the need for workforce adaptation through upskilling and reskilling.

speakers

– Mohamed Kande
– Ravi Kumar S.

arguments

AI can increase productivity and create new job opportunities

Upskilling and reskilling are crucial for workforce adaptation

Both speakers highlight the importance of involving workers in AI implementation and recognizing their potential to contribute to and benefit from AI technologies.

speakers

– Luc Triangle
– May Habib

arguments

Workers want to be involved in AI implementation processes

AI can enable non-technical workers to become builders

Unexpected Consensus

Regulation of AI output rather than input

speakers

– Ravi Kumar S.
– Luc Triangle

arguments

Regulating AI output may be more effective than regulating input

Human oversight and control are necessary in AI implementation

explanation

Despite coming from different perspectives (business and labor), both speakers agree on the need for some form of regulation or control over AI, with a focus on its output and implementation rather than its development.

Overall Assessment

Summary

The main areas of agreement include the importance of AI access, the need for upskilling and reskilling, the potential of AI to create new opportunities, and the necessity of some form of regulation or oversight in AI implementation.

Consensus level

There is a moderate to high level of consensus among the speakers on the broad implications of AI for the workforce and society. This consensus suggests a shared understanding of the challenges and opportunities presented by AI, which could facilitate more coordinated efforts in policy-making and implementation strategies for AI adoption across different sectors and stakeholders.

Different Viewpoints

Approach to AI regulation

speakers

– Ravi Kumar S.
– May Habib

arguments

Regulating AI output may be more effective than regulating input

Deregulated environments may accelerate AI adoption in companies

summary

Ravi Kumar S. suggests regulating AI output for responsible use while allowing innovation, whereas May Habib implies that deregulation accelerates AI adoption in companies.

Unexpected Differences

Role of government in AI development

speakers

– Mohamed Kande
– Ravi Kumar S.

arguments

Access to AI technology is key for workforce augmentation

Regulating AI output may be more effective than regulating input

explanation

While not explicitly disagreeing, Kande’s emphasis on access and Kumar’s focus on regulation reveal an unexpected difference in their views on the role of government in AI development. Kande implies a more active government role in ensuring access, while Kumar suggests a more hands-off approach focused on output regulation.

Overall Assessment

summary

The main areas of disagreement revolve around the approach to AI regulation, the role of government in AI development, and the balance between promoting AI adoption and addressing potential inequalities.

difference_level

The level of disagreement among the speakers is moderate. While there are some differences in approach and emphasis, there is a general consensus on the importance of AI adoption and the need to address its impacts on the workforce. These differences highlight the complexity of managing AI’s integration into the workforce and economy, suggesting that a multifaceted approach involving various stakeholders may be necessary for effective AI governance and implementation.

Partial Agreements

All speakers agree on the importance of access to AI technology, but they differ in their focus. Kande and Kumar emphasize the benefits of access for workforce augmentation and job opportunities, while Triangle warns about the potential to exacerbate the global digital divide if access is not implemented inclusively.

speakers

– Mohamed Kande
– Ravi Kumar S.
– Luc Triangle

arguments

Access to AI technology is key for workforce augmentation

AI can democratize knowledge and lower entry barriers to jobs

AI may exacerbate global digital divide if not implemented inclusively

Similar Viewpoints

Both speakers view AI as a tool for increasing productivity and creating new job opportunities, emphasizing the need for workforce adaptation through upskilling and reskilling.

speakers

– Mohamed Kande
– Ravi Kumar S.

arguments

AI can increase productivity and create new job opportunities

Upskilling and reskilling are crucial for workforce adaptation

Both speakers highlight the importance of involving workers in AI implementation and recognizing their potential to contribute to and benefit from AI technologies.

speakers

– Luc Triangle
– May Habib

arguments

Workers want to be involved in AI implementation processes

AI can enable non-technical workers to become builders

Key Takeaways

AI adoption is happening rapidly, especially among consumers, but businesses are lagging behind in implementation

Upskilling and reskilling workers is crucial for successful AI integration and workforce adaptation

AI has the potential to democratize knowledge, lower entry barriers to jobs, and increase productivity

There are concerns about AI exacerbating global inequalities if not implemented inclusively

Human oversight and control are necessary in AI implementation to mitigate risks

The AI ecosystem can create jobs beyond just the technology sector, potentially benefiting developing countries

Regulating AI output may be more effective than regulating input, but this presents challenges

Resolutions and Action Items

Companies should prioritize providing AI literacy training and tools to all employees

Organizations should involve workers in AI implementation processes

Efforts should be made to ensure widespread access to AI technology across different regions and socioeconomic groups

A new social contract may be needed to address AI’s impact on work and society

Unresolved Issues

How to effectively regulate AI given its rapid development and private sector-led innovation

How to bridge the global digital divide in AI adoption and benefits

The timeline for widespread AI deployment in mid-sized companies

How to ensure productivity gains from AI translate to wage increases for workers

Suggested Compromises

Balancing innovation with regulation by focusing on regulating AI outputs rather than inputs

Combining human oversight with AI augmentation to maintain control while benefiting from increased productivity

Investing in AI infrastructure (e.g., data centers) in developing countries to promote global participation in the AI ecosystem

This is a unique technology. Going back to what you said, it’s going to disrupt a knowledge worker, it’s going to disrupt workers who do repetitive tasks. And we’ve never had a situation of that kind. So it’s a genuine opportunity for us to disrupt the work template, the template which came from the industrial revolution.

speaker

Ravi Kumar S.

reason

This comment is insightful because it frames AI as a transformative technology that could fundamentally reshape how work is organized, challenging long-standing paradigms.

impact

It shifted the conversation to consider broader, systemic changes AI could bring about, beyond just job displacement or augmentation.

We have a unique opportunity to build that bridge, create that upward social mobility using this tooling. If we pivot this well, don’t fall into the Alan Turing trap, as I spoke about, and create shared prosperity.

speaker

Ravi Kumar S.

reason

This comment is thought-provoking as it presents AI as a potential equalizer that could create new opportunities for social mobility, rather than just a threat to jobs.

impact

It introduced a more optimistic perspective on AI’s potential societal impacts, leading to discussion of how to ensure benefits are broadly shared.

Workers want to be part of the process in their companies, in their sectors, in their countries, because there are risks. Let’s face the reality as well. There are jobs disappearing currently already, but there is also an augmentation in the jobs.

speaker

Luc Triangle

reason

This comment brings attention to the importance of including workers in the AI implementation process and acknowledges both the risks and opportunities.

impact

It shifted the discussion towards the need for inclusive approaches to AI adoption and the importance of worker involvement.

AI is very different from what we had with the Internet. The Internet was digital, the world was digital. AI is digital and physical at the same time, that’s a big difference.

speaker

Mohamed Kande

reason

This insight highlights a crucial distinction between AI and previous technological revolutions, emphasizing its broader impact across both digital and physical domains.

impact

It expanded the conversation to consider the wider ecosystem effects of AI, including job creation in various sectors beyond just technology.

What percentage of your company has built an app or an agent themselves personally? I would say access and re-skilling.

speaker

May Habib

reason

This comment cuts to the heart of practical AI implementation, emphasizing the importance of hands-on experience and skill development.

impact

It steered the conversation towards concrete measures for AI adoption and the importance of widespread participation in AI development within organizations.

Overall Assessment

These key comments shaped the discussion by broadening its scope from immediate concerns about job displacement to considering AI’s potential for systemic change in work organization, social mobility, and economic structures. They highlighted the need for inclusive approaches to AI adoption, involving workers and considering global implications. The conversation evolved from focusing on challenges to exploring opportunities for shared prosperity and the unique characteristics of AI that differentiate it from previous technological revolutions. This led to a more nuanced and forward-looking dialogue about how to harness AI’s potential while addressing its risks.

How can governments be involved in AI workforce development without hindering private sector innovation?

speaker

Mina Al-Oraibi

explanation

This question addresses the balance between government regulation and private sector leadership in AI adoption and workforce transformation.

How can companies implement change management strategies to help employees adapt to AI?

speaker

May Habib

explanation

This area explores practical approaches for companies to support their workforce in transitioning to AI-augmented work environments.

How can we ensure that AI adoption doesn’t exacerbate existing inequalities between developed and developing nations?

speaker

Luc Triangle

explanation

This question addresses concerns about the global digital divide and ensuring equitable access to AI technologies across different regions.

What strategies can be employed to create upward social mobility in jobs through AI augmentation?

speaker

Ravi Kumar S.

explanation

This area of research focuses on using AI to create opportunities for workers to advance in their careers and increase their earning potential.

How can we ensure that productivity gains from AI translate into wage increases for workers?

speaker

Luc Triangle

explanation

This question addresses concerns about ensuring fair distribution of the economic benefits of AI adoption.

What is the timeline for widespread AI adoption among mid-size companies?

speaker

Audience member (Christy Hoffman)

explanation

This question seeks to understand the expected pace of AI deployment beyond large corporations.

How can we effectively regulate AI outputs rather than inputs?

speaker

Ravi Kumar S.

explanation

This area of research explores potential approaches to AI regulation that balance innovation with responsible use.

What metrics can be used to measure access to AI technologies?

speaker

Mohamed Kande

explanation

This question seeks to establish concrete ways to assess and ensure equitable access to AI tools across populations.