Corporate Ladders, AI Reshuffled
20 Jan 2026 16:30h - 17:15h
Corporate Ladders, AI Reshuffled
Session at a glance
Summary
This Davos 2026 panel discussion focused on how artificial intelligence is reshaping the job market and corporate workforce structures, examining both the disruption and opportunities AI presents for entry-level workers and beyond. The conversation featured perspectives from education, economics, technology, and business leaders who debated whether AI will cause catastrophic job displacement or create new opportunities requiring different skills.
Andrew Ng argued that while AI headlines suggest widespread job destruction, the reality is more nuanced, with AI capable of handling 30-40% of most jobs while humans remain essential for the remaining tasks. He emphasized that workers who know how to use AI will replace those who don’t, and advocated for teaching everyone to code as AI tools make programming more accessible across all professions. Professor Christopher Pissarides countered that many jobs in healthcare, hospitality, and retail require human empathy and interpersonal skills that AI cannot replicate, suggesting the disruption may be less severe than predicted.
UAE Education Minister Sarah bint Yousif Al Amiri highlighted her country’s pioneering approach as the first nation to mandate AI literacy in K-12 education, emphasizing the importance of teaching students not just how to use AI tools, but when to use them and how to think critically about AI-generated content. The panel agreed that the education system must shift from knowledge transmission to teaching students how to learn, unlearn, and relearn continuously.
Himanshu Palsule noted the irony that while companies worry about AI displacing workers, significant skill gaps persist, particularly in technical roles that require AI proficiency. The discussion concluded that success in an AI-enabled future requires combining broad AI literacy with deep domain expertise, continuous learning capabilities, and strong collaboration between educational institutions, governments, and private sector employers.
Keypoints
Major Discussion Points:
– AI’s Impact on Entry-Level Jobs and the Skills Gap: The panel discussed whether AI will cause catastrophic job displacement or create new opportunities, with emphasis on the paradox that while some fear mass unemployment, there are significant skill shortages in sectors like healthcare (11 million gap) and technology.
– Education System Reform and AI Literacy: A central focus on how educational institutions must adapt, highlighted by the UAE’s pioneering K-12 AI literacy mandate for 280,000 students, and the need for universities to prepare students for AI-integrated jobs rather than pre-AI era positions.
– The Universal Need for Coding Skills: Andrew Ng’s controversial assertion that “everyone needs to learn to code” in the AI era, not traditional programming but AI-assisted coding, as non-technical professionals (marketers, HR, finance) who can code are becoming significantly more productive.
– Workforce Transformation and Corporate Strategy: Discussion of how companies need to move beyond bottom-up AI point solutions to top-down business process redesign, and how the workforce structure will change as humans work alongside AI agents and digital labor.
– Societal and Policy Implications: The moral obligation to retrain displaced workers, the need for public-private partnerships in education, and addressing public mistrust of AI while managing geopolitical and economic disruptions that compound AI’s impact.
Overall Purpose:
The discussion aimed to examine the real-world implications of AI on employment, moving beyond sensationalized headlines to explore practical solutions for workforce adaptation, educational reform, and corporate strategy in an AI-driven economy.
Overall Tone:
The tone was notably optimistic and solution-oriented rather than alarmist. While acknowledging legitimate concerns about job displacement, panelists consistently emphasized opportunities over threats. The conversation maintained a collaborative, expert-level discourse with occasional respectful disagreements that enriched the debate. The tone remained constructive throughout, focusing on actionable strategies rather than dwelling on dystopian scenarios, with panelists building on each other’s insights to present a balanced view of AI’s workforce impact.
Speakers
– Shereen Bhan – Managing Editor of CNBC TV18, CNBC’s operations in India, Session Moderator
– Sarah bint Yousif Al Amiri – Minister of Education of the United Arab Emirates
– Christopher Pissarides – Regius Professor of Economics at the London School of Economics and Political Science from the UK
– Andrew Ng – Founder of deeplearning.ai from the US, Chairman of Coursera
– Himanshu Palsule – Chief Executive Officer of Cornerstone on Demand from the US
– Audience – Various audience members asking questions during the session
Additional speakers:
– Derby Chukwudi – Global Shaper based in Hong Kong
– Kian – Founder and CEO of Orkera
– Unnamed audience member – Founder of Waffle (addressing gender gap industry in Japan), engages in policy recommendation in Japan
– Unnamed audience member – Works on assessing AI’s impact on the workforce
Full session report
Comprehensive Report: AI’s Impact on the Future of Work – Davos 2026 Panel Discussion
Executive Summary
This Davos 2026 panel discussion brought together leading voices from education, economics, technology, and business to examine how artificial intelligence is fundamentally reshaping the job market and corporate workforce structures. Moderated by Shereen Bhan, Managing Editor of CNBC TV18, the session featured Sarah bint Yousif Al Amiri, Minister of Education of the United Arab Emirates, Christopher Pissarides from the London School of Economics, Andrew Ng, founder of deeplearning.ai and chairman of Coursera, and Himanshu Palsule, CEO of Cornerstone on Demand.
The discussion moved beyond sensationalised headlines about AI-driven job displacement to explore practical solutions for workforce adaptation, educational reform, and corporate strategy. The tone remained notably optimistic and solution-oriented, with panellists consistently emphasising opportunities over threats whilst acknowledging legitimate concerns about disruption. The conversation featured active audience participation and revealed both significant consensus on key issues and meaningful disagreements that enriched the debate.
Major Discussion Points and Speaker Positions
AI’s Impact on Employment and Job Transformation
The panel addressed widespread anxiety about AI’s potential to cause catastrophic job displacement, with speakers offering nuanced perspectives that challenged simplistic narratives.
Andrew Ng argued that whilst AI headlines suggest widespread job destruction, the reality is more complex. He emphasized that “people using AI will replace those who don’t” rather than AI replacing people entirely. Ng made a definitive statement about his hiring practices: “I’m not going to ever hire another engineer again that doesn’t know how to use AI to help them code.” However, he acknowledged that certain roles face complete displacement, specifically mentioning translators, voice actors, and call centre operators as particularly vulnerable.
Ng also provided important context for current job market disruptions, arguing that recent layoffs are primarily due to pandemic over-hiring rather than AI displacement. This perspective challenged the fundamental premise driving much of the public anxiety about AI’s immediate impact on employment.
Christopher Pissarides offered a more reassuring view, arguing that approximately half the workforce—particularly in healthcare, hospitality, and retail—won’t be significantly threatened by AI because these roles require human qualities like empathy and interpersonal skills that AI cannot replicate. He noted that “most jobs involve tasks AI cannot fully automate” and emphasised the irreplaceable value of human customer interaction in service industries.
However, Pissarides identified professional services as facing the greatest disruption, particularly law and accounting. He provocatively suggested that these professions “were structured more or less by the British in the 19th century” and that “it’s about time they got a kick in the backside and they reformed.”
Himanshu Palsule highlighted a critical paradox in current workforce discussions: whilst companies worry about AI displacing workers, significant skill gaps persist, particularly in technical roles requiring AI proficiency. He noted the irony that represents “the greatest irony of our time” – that organizations are “leaving behind a generation that is actually the most capable of implementing AI” whilst focusing resources on re-skilling middle and senior management.
Education System Reform and AI Literacy
The discussion revealed strong consensus that current educational systems are fundamentally inadequate for preparing students for an AI-integrated economy.
Sarah bint Yousif Al Amiri presented the UAE’s pioneering approach as the first country to mandate AI literacy in K-12 education for all students. She emphasised that their programme goes beyond teaching students how to use AI tools to include when to use them and how to think critically about AI-generated content. Drawing parallels with social media adoption, she warned: “We looked at social media. Students have used social media extensively, haven’t been given a code of ethics on how to use it effectively… we’ve seen adverse effects on both their personalities, their social well-being, and their social interactions as well.”
The UAE’s approach includes teaching students to evaluate AI outputs for bias and accuracy, recognising that unguided technology adoption can have profound developmental consequences. Al Amiri stressed that education must focus on teaching students “how to learn, relearn, and unlearn continuously” rather than transmitting static knowledge.
Andrew Ng was particularly critical of higher education, stating: “Unfortunately, I feel like higher education, which I love and I think is a great force for the good, is failing many fresh college graduates by preparing them for the jobs of 2022 before modern AI, rather than the jobs of 2026 and beyond.”
Christopher Pissarides advocated for extending university education and increasing industry collaboration through internships and apprenticeships. He shared a personal example about his son learning both Latin and programming in English schools, suggesting that diverse skill development remains valuable even in an AI-driven world.
The Universal Coding Debate and Audience Interaction
One of the most engaging segments emerged around Andrew Ng’s assertion that “everyone needs to learn to code” in the AI era, which sparked significant audience participation.
Ng’s position was that AI-assisted coding is becoming as fundamental as literacy across all professions. He provided concrete examples: “We’re already seeing in Silicon Valley that not just the software engineers, but the marketers, HR professionals, financial analysts, and so on, the ones that know how to code are much more productive than the ones that don’t, and that gap is growing.”
This assertion prompted a thoughtful challenge from a Global Shaper from Hong Kong in the audience, who questioned whether universal coding education might undermine economic principles of comparative advantage and specialisation. The audience member asked whether this approach conflicts with the economic benefits of having people specialize in different areas.
Christopher Pissarides responded by comparing coding education to learning foreign languages—useful for familiarity and understanding but not expecting everyone to become expert programmers. He supported the concept whilst acknowledging practical limitations.
Himanshu Palsule reinforced Ng’s perspective with a personal example about his son using Python for cognitive science research, illustrating how coding skills enhance capabilities across disciplines. He also noted that computer science graduates are “waiting for big tech jobs that no longer exist whilst other departments need AI skills.”
Workforce Transformation and Corporate Strategy
The discussion revealed significant insights about how companies must approach AI integration and workforce development.
Andrew Ng distinguished between incremental and transformative AI implementation, providing a detailed example of loan underwriting workflow redesign. He explained how traditional loan processing involves sequential steps—application review, credit check, income verification—that could be completely reimagined with AI to create parallel, more efficient processes. This illustrated his broader point about moving “beyond bottom-up point solutions to top-down business process redesign.”
Ng emphasized that whilst efficiency gains from AI point solutions might yield modest improvements, transformative workflow redesign could deliver much greater value. However, he acknowledged that “workflow redesign is difficult but necessary—requires rethinking entire value creation processes rather than optimising individual steps.”
Himanshu Palsule highlighted fundamental challenges in human capital policies, noting that “work output now includes human plus AI agent collaboration” and that traditional HR frameworks are inadequate for this new reality. He emphasized that companies should focus on developing “judgment, decision-making, and context skills rather than competing with AI on repetitive tasks.”
The moderator, Shereen Bhan, noted in her opening remarks that there is an 11 million healthcare worker gap globally, highlighting how skills shortages persist even amid concerns about AI displacement.
Societal Trust and Public Acceptance
A particularly powerful moment came when Andrew Ng shared a personal anecdote that brought sobering reality to the elite discussion: “About two weeks ago, my team… spoke of a coffee shop owner in part of Silicon Valley… who was politically shaking, because he was so angry at AI… I think many people underestimate the degree to which our sector is mistrusted, and sometimes even hated.”
This story illustrated the disconnect between optimistic AI discourse at forums like Davos and actual public sentiment, introducing crucial dimensions of social acceptance and trust that could ultimately determine AI adoption success.
Sarah bint Yousif Al Amiri reinforced the need for responsible implementation, arguing that “government has obligation to reskill workers and foot the bill for retraining programmes.”
Christopher Pissarides noted additional complexity from “geopolitical fragmentation and deglobalisation” that compounds AI workforce disruption, whilst Himanshu Palsule highlighted the need for “cross-border skills mobility as talent gaps exist globally whilst policies become more protectionist.”
Areas of Consensus and Disagreement
Strong Areas of Agreement
The discussion revealed remarkable consensus across several key areas:
Education System Reform: All speakers agreed that current education systems require fundamental reform from K-12 through university levels, with educational institutions failing to prepare students with relevant AI skills needed for the modern economy.
Human-AI Collaboration Model: There was strong agreement that most jobs will involve human-AI collaboration rather than complete human replacement, with humans focusing on tasks requiring judgment, empathy, and contextual understanding.
Continuous Learning Imperative: All speakers emphasized that the ability to continuously learn and adapt is more important than acquiring static knowledge, given the pace of technological change.
Societal Responsibility: Clear agreement emerged that society has a collective responsibility to support workers displaced by AI, with both Ng and Al Amiri explicitly stating moral obligations for retraining and support.
Key Disagreements
Scope of AI Job Disruption: Christopher Pissarides argued that approximately half the workforce in service sectors won’t be significantly threatened, while Himanshu Palsule referenced Yuval Harari’s broader perspective that any job involving words, language, and numbers will face disruption.
Universal Coding Education: While there was general support, the audience challenge about comparative advantage highlighted tensions about whether universal coding conflicts with economic principles of specialization.
Primary Cause of Current Disruptions: Andrew Ng’s argument that current layoffs stem from pandemic over-hiring rather than AI displacement challenged the fundamental urgency driving the discussion.
Thought-Provoking Insights
Several comments provided particularly profound insights that shaped the conversation’s direction:
Andrew Ng’s hiring declaration fundamentally reframed the discussion from ‘AI will eliminate jobs’ to ‘AI illiteracy will eliminate job prospects,’ establishing a solutions-oriented tone.
Sarah bint Yousif Al Amiri’s social media analogy provided crucial historical context about the importance of proactive education rather than reactive regulation when adopting transformative technologies.
Christopher Pissarides’ historical perspective on professional services being “structured by the British in the 19th century” reframed disruption as potentially beneficial reform rather than destructive change.
The coffee shop owner anecdote powerfully illustrated the trust gap between AI optimists and the broader public, highlighting a critical challenge for successful AI adoption.
Unresolved Issues and Future Directions
The discussion highlighted several areas requiring further investigation:
Educational Framework Development: While there was consensus on the need for AI literacy, specific questions remain about optimal curriculum depth, progression across educational levels, and effective evaluation methods.
Workforce Transition Management: Key unresolved issues include systematic identification and support for workers in highly automatable roles, redesign of human capital policies for human-AI collaboration, and management of transition periods.
Trust and Acceptance: The discussion raised but didn’t resolve how to build societal trust in AI and bridge the gap between elite optimism and public skepticism.
Recommendations and Action Items
Based on the discussion, several clear action items emerged:
For Governments: Mandate AI literacy in K-12 education following the UAE model, fund comprehensive retraining programmes for displaced workers, and develop frameworks for cross-border skills recognition.
For Educational Institutions: Reform curricula to include AI skills across disciplines, increase industry partnerships, and teach continuous learning capabilities rather than static knowledge.
For Businesses: Invest in transformative workflow redesign rather than just efficiency improvements, create opportunities for AI-fluent young workers, and focus hiring on AI capabilities.
For Individuals: Develop broad AI literacy combined with deep domain expertise, focus on judgment and decision-making skills, and cultivate continuous learning capabilities.
Conclusion
This Davos 2026 panel discussion represented a mature examination of AI’s impact on work that moved beyond simplistic job displacement fears to comprehensive workforce adaptation strategies. The strong consensus on education reform, human-AI collaboration, and societal responsibility suggests potential for coordinated policy responses.
However, the disagreements about disruption scope, universal coding education, and current market causes highlight the complexity ahead. The discussion’s most valuable contribution may be its reframing from defensive job protection to proactive skills development and strategic workforce planning.
The emphasis on public trust, illustrated by the coffee shop owner anecdote, suggests recognition that technical solutions alone are insufficient. Successful navigation of AI’s workforce impact will require coordinated efforts across institutions to ensure AI benefits are broadly shared whilst supporting those facing displacement.
As the moderator noted in closing, this remains an “ongoing debate” requiring continued dialogue between technology leaders, policymakers, educators, and the broader public to shape an inclusive AI-enabled future of work.
Session transcript
meeting in Davos 2026. I’m Shirin Bhan, the moderator for this session, also the managing editor of CNBC TV18, CNBC’s operations in India. It’s great to see so many wonderful people here, and it’s great to have a panel here to discuss a topic that is getting a lot of attention and mind space here in Davos.
You know, we’re talking about corporate ladders, AI reshuffle, but in the context of Davos 2026, reshuffled has taken on a whole new meaning. We’re, of course, waiting for President Trump to arrive, and it could take on an entirely different meaning here in the context of what happens with headlines. But I want to dive into the issue at hand.
I think there is palpable anxiety about what AI is going to do in the near term, in the medium term, and in the long term to the jobs market. If you take a look at the World Economic Forum’s jobs report, while there will be disruption, it does say that there will be net growth as far as jobs are concerned, at least at this point in time. The irony also is that while we talk about a catastrophic impact on entry-level jobs, there are sectors where we continue to see significant gaps.
There aren’t enough skilled workers. The WEF in its report also talks about sectors like healthcare, where the gap is almost 11 million. You don’t have 11 million people to service the healthcare needs.
So that is also the irony that we are dealing with. But there’s a lot of questions on what companies need to do, what companies are doing today, and whether that is adequate to upskill workers. What happens to entry-level workers?
Do we need software engineers anymore, or is code going to be written by large language models and generative AI? I think that’s what we intend to dive into. This is going to be a free-flowing conversation.
I would appreciate it. If you do have questions, then do engage with our panel here. Do keep your questions short.
We will come to you in just a second. But let’s get things started by introducing our panellists to you, Sarah Bind Youssef Al-Amiri, the Minister of Education of the United Arab Emirates. Thank you very much for joining us.
Christopher Biserradi is the Regius Professor of Economics at the London School of Economics and Political Science from the UK. Andrew Ng, founder of deeplearning.ai from the US. Himanshu Palsula, the Chief Executive Officer of Cornerstone on Demand, also from the US.
Thank you very much for joining us here. Andrew, let’s get started. I think a lot of the headlines scream catastrophic disruption.
What is going to happen as far as entry-level jobs are concerned? Will it be a white-collar bloodbath, so to speak? Let me get
your view on that to start with. I think the entry-level jobs are there, and many businesses just can’t find enough skilled entry-level workers with the right AI skills. Unfortunately, I feel like higher education, which I love and I think is a great force for the good, is failing many fresh college graduates by preparing them for the jobs of 2022 before modern AI, rather than the jobs of 2026 and beyond.
We’re already seeing that in coding, for example, I’m not going to ever hire another engineer again that doesn’t know how to use AI to help them code. Because AI tools have advanced so rapidly in software engineering and coding, coding is, I think, a harbinger of what will happen in other sectors as well, as tools enter other sectors. And I feel a lot of urgency to revise the academic curricula to give the students they need to be job-ready, because the jobs are there.
The jobs are there, the skills gap continues
to exist, and it is only going to widen if we don’t do something about the way that our education system functions. That’s the point you’re making? Absolutely.
Minister, let me come to you now with that, because that’s exactly the problem that you’re trying to address at the UAE, and you’ve actually brought in AI as part of the curriculum K-12. Yes, absolutely. So that’s a very interesting
point when it comes to higher education, which is the skilled workforce. I’ll address that first, and then take it back, because we have a bit more time in preparing people to come to the workforce when it comes to K-12. When we’re talking about higher education, higher education needs to start infusing AI, and we do have that in the UAE, starting to work with some of our universities to provide them with skills and the necessary tool sets to be able to advance with that.
Right now, it’s starting with basic AI literacy in higher education, but that needs to improve in subject level. So based on the subject level that each student in higher education is currently going through and the degree that they’re pursuing, there needs to be a better tackling on the tools that today exist in the market and the rapid evolution of the tools.
Now, how are we taking it to K-12? We’re the first public school system globally to mandate K-12 AI literacy for all of our students. So today, we have more than 280,000 students taking AI literacy at least once or once every two weeks.
The core of the curriculum is the following. Students need to understand how to use AI, when to use it, what tools to use to be able to support them in their learning, and to be able to advance that. We’ve also instilled new training for educators on how to use existing AI tools to ease their job and be able to do their job more effectively and more efficiently and free up some of their time that is used on day-to-day work.
We also need to relook the outcomes of the student themselves, and that’s more about the broader long-term impact and changes that need to happen in the education system. You know, that’s very interesting what you point out, that the UAE is the first country in the world to mandate AI curriculum, and why did you feel the need to do that? We looked at social media.
Students have used social media extensively, haven’t been given a code of ethics on how to use it effectively, what are the challenges of not using it in the correct mechanisms, and we’ve seen adverse effects on both their personalities, their social well-being, and their social interactions as well.
We saw that that might happen in their cognitive ability, the development of their cognitive ability when it comes to AI, so it was very important for us to design a curriculum for them to understand how to use it.
We’ve embedded subtly critical thinking also within the realms of AI, so the class goes as follows. A student uses AI, for example, let’s talk about prompt engineering, prompting it to get information about a particular subject, let’s say World War I, and they try different prompts, and the teacher speaks with them based on critical thinking on what prompt worked better, what information was not biased, what information came out in accordance to what you understand, and that’s also instilling a bit of change that we need to see in the education system, which is it’s no longer about providing the knowledge and information to students and asking them to repeat it back, it’s more about there’s a lot of information and knowledge out there, can you pick out what is relevant, can you understand what is construed or skewed, and are you able to bring together groups and pieces of information
and knowledge to be able to get a proper outcome and analyze that properly? Yeah, Andrew, before I get Himanshu and Professor Pesirade into the conversation, I just want to go back to the point that you were making, because in a way, what we heard there from the Minister addresses the issue that you were raising, do you believe that this needs to be the way forward across
different countries? I think K-12 is an excellent time to start training people up on AI, maybe in terms of the AI disruption of jobs at the entry level and beyond, I think the hype is beyond the reality, there have been layoffs, so on over the past year or two, from where I’m sitting, a lot of it seems to be from overhiring from the pandemic rather than AI-affected, but looking to the future, I know Chris was an economist, a lot of my economist friends, everyone else and others, have done these task-based analysis of jobs, where you take a job, break it down into tasks, and figure out what AI can or cannot do.
It turns out for many jobs, maybe AI could do 30-40% of the job, that means we still need people to do that 60-70%, but it is also true that someone that knows how to use AI will replace someone that doesn’t, even if AI itself won’t replace a person, so getting through the hype to give people the skills they need is critical, and I would just add one asterisk to what I said, which is, whereas for most jobs you still need people to do 60-70%, there’s a small minority of jobs where AI can indeed do almost 100% of it, so those jobs will run into trouble.
I’m worried about the translators, I’m worried about the voice actors, I’m worried about the call centre operators, and for those people, we actually have an obligation, I think, as a society, it’s a moral obligation to take care of them, because just because someone’s job is going away does not mean they deserve to be thrown onto the street, and so the education system, up-schooling, I feel like we need to get that right as well.
I think the moral obligation
that you speak of is an important but a complex issue that will have to be dealt with contextually in different jurisdictions, but Professor, let me come to you on the hype versus reality, and we see these screaming headlines of so many jobs gone on account of AI.
To Andrew’s point, a lot of this could potentially have been because of over-hiring through the past… especially as far as the tech companies are concerned. You’ve been speaking with tech CEOs today.
What are they telling you about entry-level jobs?
Yeah, I’m afraid Andrew has stolen my thunder now. That’s exactly what I was going to say. You know what I’m saying?
Is it like… I mean, I was going to say… I don’t want to spoil the Davos party, where everyone is saying, you know, if you look at it carefully, if you look at the labour market carefully, what do people do?
Young people, if you like entry-level, there are entry-level people who are not young. About half, I think Andrew said 40%, well, we’ll take 40%. We’ll never use AI.
They don’t need AI now. They go into health and care, hospitality, retail. These are usually not university graduates, though, so I don’t want to…
Minister, I’ll… I mean, they’re wasting their time training because she was obviously talking about universities. So, half the workforce is out.
Of what remains, most of them will need to use AI, but their jobs are not threatening. AI cannot do what the job will be doing. In contrast to that, what they will be required to do by their companies is that they will apply some AI skills that they have and the company will give them AI, will invest in AI, and they will become more productive.
They will do things faster, better quality, and everything like that. Now, those people are usually university graduates and they’re the ones that are benefiting from university education in AI and I think at LSE, we’re moving in that direction too. We’ve been talking to Anthropic and so on.
Then what remains are the people that if not 100%, like Andrew said, maybe it’s 80% that they can do it and usually those, and the reason you get to hear so much about those is that they’re usually in the professions.
It’s the way, you know, if you look at law, accounting, I mean, those professions were structured more or less by the British in the 19th century mainly and they did it according to the education that existed at the time.
You know, if you have usually a son, but let’s say a child, a son or a daughter, what would they do? They would go to a private, what they call public school, but it’s such a private boarding school. They will come out.
They will go to Oxford and Cambridge. They will get classics education like Boris Johnson or something similar and then they will enter one of the professions, law, accounting, where they will get their privilege and their training and gradually they become partners.
That’s the one that is hit most because that’s what generative AI and the large language models can do and in fact, it’s about time they got a kick in the backside and they reformed the professions. I mean, you know, you cannot leave 19th century English structure of education and professions.
To borrow from Professor Pissarides very eloquently, who is going to get a kick in the backside?
So, Shirin, you talked about ironies. I think the greatest irony of our time is we are leaving behind a generation that is actually the most capable of implementing AI and we are spending time re-skilling middle management and senior management, which is harder.
So, I think, Minister, the model that you have, I would say absolutely every school needs to follow that model. I think when you get junior people at work, I think it will be a mistake to introduce them to autonomous repeatable roles because they’re going to go away. I would focus on skills that involve judgment, decision-making, context because those are going to be more important.
I would rather have a junior financial analyst or a legal analyst coming in and spend their time examining the output of an AI tool rather than competing with AI. So, as we look at these job structures, I believe what these young adults coming in, our kids, our grandkids, need to sort of look at this T-shaped model where the breadth is get good enough with AI and the depth is pick a domain and then get specialized in that domain because that is going to become irreplaceable.
And then for companies hiring these people, create those environments, create those apprenticeships, create the area where they come in, especially if you have this education system K-12 where they’re AI fluent.
I think we need to trust these people who are born digitally native in being able to adopt to this. So, you know, I’m very optimistic, but I think we need a fundamental reset in not just education, but in hiring, onboarding, and training of these people.
And I want to pick up from what Professor Bessiradis just said that, you know, are we largely talking about what’s likely to happen in the tech sector or tech affiliated sectors? Will sectors like hospitality, healthcare, and so on and so forth be as impacted by what we’re speaking?
I think so. If you heard what Yuval Harari had to say on stage, he said any job that involves words, language, and numbers is going to get disrupted because words, language, and numbers, the cohort information available to an AI agent is far going to surpass our ability to compete with that.
So whether you’re a lawyer who has to read all the precedent books from the past, you’re a financial analyst who has to understand complex models, you’re a healthcare worker who has to understand efficacy of medicine, pharmaceuticals, and diseases, and all of that, don’t compete with your corpus of knowledge with what’s going to be available there.
Highlight the fact that context, judgment, and decision making is going to be more important. So I respectfully disagree that while…
No, I love the fact that there’s disagreement because that’s the whole point of the debate.
You mentioned word. I also add images because it also influences a lot the healthcare sector and diagnosis as well.
I love the fact that you’re waiting to jump in. Go right ahead.
I agree. I agree both with Harari and… Sorry, I didn’t think I would say that.
If it involves words, where the disagreement is, and I’m sure you’re going to agree when you hear it, is that these jobs that do not use AI, they are not words jobs. If you become a nurse, your skill is not that you know how to put words together. You need empathy.
You need to understand the person you are looking after. As Harari said, we have absolutely no evidence that AI has any feelings at all. If you go into the health sector, you don’t need to have any kind of feeling.
You need to have any special feelings. By the time AI gets there, the current generation of nurses will no longer be there. It will be their grandchildren.
Going to the hospitality industry with these people who serve us out here, who give us those amazing espressos and other things that they serve, is their skill putting words together? Of course not. Their skill is knowing how to treat the customer.
Here we’re not in prison. They’re a captive customer. But if they’re going to attract you to go to their restaurant or their hotel or something, they need to feel that they understand you.
They understand the customer. They’re nice. It’s not a word.
If you put an AI agent at reception and you put the really nice people next door to another hotel, which one would you go?
24 by 7, no vacation, no paid leave. I don’t know. That’s probably the CEO.
Andrew, let’s double down on the point that’s being made there. We’re talking about a world of agents. We’re talking about a world where digital labour and human labour will coexist in an environment like that.
How do corporate leaders look at staffing? How do corporate leaders look at the challenges of workforce, workforce management? You talk to CEOs.
What are they telling you at this point in time? What is the dilemma that they’re faced with?
I find the term digital labour challenging because I’m the one that coined the term agentic AI. We saw this coming quite early. Digital labour is a challenging concept because AI software is so different than humans.
They’re analogising them to needing food and water and paid leave and so on. These problematic analogies. When I speak of CEOs, I think the two biggest challenges, one is upskilling, bring people along with us, because ultimately it’s not the technology by itself.
It’s the people we bring with us that will implement the technology that drives the change. In fact, my chain, Cutting Fruits, does a lot of work on skill measuring to help people figure out the skill gaps and what they need to learn. So that’s one.
And the second thing is, to the business leaders, I find that we’ve done this let 1,000 flowers bloom bottom of innovation thing. And for the most part, it’s led to a lot of nice little things, but nothing transformative for businesses. And what I’m seeing for businesses is, to execute the more transformative projects, the problem of a lot of bottom-up innovation is, if you look at the set of tasks needed to create value in the business, a lot of bottom-up innovation ends up with point solutions that takes one step out of five, makes it more efficient.
Then you get these 5%, 10% efficiency gains, which are nice, but not transformative. But it’s usually when the bottom-up innovation meets top-down, then you can redesign the entire workflow from scratch, and that’s what’s creating more business growth. So actually, here at this week of Davos, speaking with a lot of CEOs, they’re thinking about how to go beyond the bottom-up point solutions to taking the broader view to business process redesign.
And I think that will finally realize a lot of the growth that AI has been…
But are we closer to that? Because the discussion and the debate and the conversations that I’ve still been having with people, it is still pretty much bottom-up. It’s still efficiency and optimization.
It’s not so much transformation just yet. So do we still have time before we see this kind of transformative impact for governments to be able to skill people, for corporations to figure out what they intend to do as far as the corporate structure is concerned, or are we out of time?
We still have time. Frankly, this workflow redesign is really hard. Maybe one example that’s been around for a little bit.
Take underwriting loans. There are multiple steps needed to do that. You have to do loan approval, preliminary approval, diligence, execute the loan, manage and give out the loan.
So there are multiple things needed to create value. So what some people said is, oh, why not use AI to do the preliminary loan approval, which is nice. But if you take this whole process and make one step a bit more efficient, you get this nice, small cost savings.
But what some businesses realize is if I get AI to automate this step, what it can do is build a brand new product where instead of just a cost savings, I’m going to say, I’m going to get back to you on your loan approval in 10 minutes instead of one week.
So this is a materially different product and it drives business growth. But it takes that broader view where it’s not someone automating one piece of the puzzle, but rethinking all the steps needed to create value to redesign that, that’s what’s driving growth. And frankly, this is really hard.
We spend a lot of time with CEOs. One large organization sent us almost 300 project ideas and asked my friend and I at AI Aspire to sort out which are the ones that can drive strategic value. And we find this work really difficult and really intellectually deep.
So I think it’ll be some time before businesses get to these, not the cost savings, but the significant growth opportunities.
Absolutely, before they move on to the transformation stage. But Himanshu, you know, again, I want to double down on this. What a company is telling you and how do you believe the workforce is likely to be restructured in the near term?
Yeah, so the skill shortage is very real. And it’s driven also by an unprecedented amount of automation. You have these complex dialysis machines sitting idle because nurses cannot get trained fast enough.
We have a customer that manufactures aircraft and the new avionics. There’s no maintenance staff doing that. We spent a decade worrying about data flows.
And it’s Davos, so let’s talk a little bit about cross-border. And we came up with data sovereignty, data governance. You know, we build laws and all of that.
We better start thinking about skills in the same way because there could be a quarter in Nigeria, there could be an analyst in Riyadh who has better skills than the people you may have just hired. And we need to be able to figure a way to fill those gaps quickly. That skills divide is going to keep getting wider, especially with more amount, more automation, more AI.
And, you know, it’s up to us to try to close that divide. The concern I have is as, you know, depending on, as you said, it’s going to be a fascinating week here in Davos. Depending on where people retrench to, we may take a bad problem and make it worse.
And then you have untethered AI that’s unfairly used across geographies. That is my biggest worry. As a global company, we hire people from all over the world.
Half our teams are outside the United States. And we really need to start thinking about this global skilling. We work with the Abu Al-Gharair Foundation in Dubai.
And companies like that need to come to the table to have this discussion with businesses.
Well, you know, easier mobility of talent. That is an increasingly large challenge in the world that we live in at this point in time, with policies becoming more inward and more protectionist, including those that involve the mobility of people.
But that’s a separate conversation. That’s a separate debate. You know, I want to go back to the point that Andrew made about moral obligation.
And I want to understand this from your perspective as somebody in government, Minister. The moral obligation, of course, is to skill, to educate. But in a situation where jobs are not probably growing fast enough, or people who may have been left behind on account of the disruptions in specific sectors, what is then the obligation of the government in that situation?
It is the obligation of government to skill. And we do have a program that was launched at the back of COVID to skill certain individuals within society to be able to start filling in jobs within our private sector and ensuring that the private sector opens up new jobs to people to be able to fit into it.
So the government is footing the bill to re-skilling people. There’s another part of it. And I’ll go back to my portfolio within the K-12 education system.
Governments have to also build a future workforce that can re-skill intrinsically. So teaching them the ability to learn and relearn and unlearn and continuous learning that we’ve been putting it as an objective for decades now, as an outcome of the education system needs to be a top priority.
Otherwise, governments who would need to continuously re-skill people will be footing the bill more and more and more down the line if it’s not intrinsically built into the workforce.
You know, Professor Bissiradis, while we talk a lot about AI and the impact it’s likely to have as far as the jobs market is concerned, I don’t think that we can see it in isolation. We’ll have to see it in the overall economic context where we are seeing geopolitical risks. We’re seeing geoeconomic confrontation, geoeconomic fragmentation.
There’s a retreat from globalization as we’ve come to accept it, as we’ve come to know it. That, in addition to AI, are we headed into a structurally disruptive environment which will change the way that the labor market operates?
Well, I’m afraid we are. It’s an unusual circumstance that we’re facing now because we have the globalization shock and we have the deterioration in geopolitics, which obviously affects the location of companies. So in addition to companies now needing to know how am I going to produce something and what skills do I need, they also need to worry about where am I going to produce this, what kind of supply chains am I going to use?
And that’s a very difficult balance to bring up. That’s very much a company’s problem now. Of course, it would not be good for their productivity or for world economic growth if each of them says, OK, I’m going to get my home, retrench, stick home and all that, which is – I don’t know if you’ve been to Macron’s talk a bit earlier.
He was talking very much about Europe needs to get together, get deeper, integrate more to the single market, single capital market. It’s all because of that. Can we trust anyone outside Europe given what we are seeing behaviors east and west from where we are?
It’s a very, very difficult problem. From the workers’ point of view, they can never lose if they do exactly what the minister said. You have to learn how to learn in future.
Don’t think that you are going to acquire a skill at school and maybe a little bit of training later, and then you’re going to get into a job and you’re always going to do that skill. I mean, it’s partly universities that need to reform, actually, so I can be – I have I have no more power to reform my university. But essentially, what they need to do is to teach a variety of skills to their students in addition.
But maybe extend, like we have three years of university undergraduate education in Britain, extend it to four maybe, and use it in combination with companies, with industry. So you have internships, or you might call them internships, or apprenticeship training. But the idea is that the university learns from the industry what skills are needed and teaches different skills.
So when they graduate, they have a composition of skills to show in their portfolio. They have their degree, obviously. And they have experience.
They have work experience. That was the most common complaint that you hear now from young people, graduates coming out of school, is that everyone wants experience, but they don’t offer me a job because they don’t have experience.
How am I going to acquire the experience if they don’t offer me a job? Get your experience at university.
And Andrew, I want you to come in on that and also build on the need and the possibility for a much more robust public-private partnership, a collaboration between industry, academia, and to be able to address the skills gap that we speak of, but also identify the skills for the jobs of the future.
Yeah, I think one of the challenges is, because AI technology is still evolving rapidly, the skills that are going to be needed in the future are not yet clear today, hence lifelong learning, the ability to keep on learning those skills in the future that we do not yet know today.
But there’s one skill that is already emerging that’s very clear to me that I want to mention that may be controversial, which is, it’s time to get everyone to learn to code. And what we’re already seeing in Silicon Valley is that you shouldn’t code the old way. You know, don’t write code by hand, get AI to do it for you.
But we’re already seeing in Silicon Valley that not just the software engineers, but the marketers, HR professionals, financial analysts, and so on, the ones that know how to code are much more productive than the ones that don’t, and that gap is growing.
So my best recruiters, they don’t read resumes by hand, they write code to screen resumes for them. When one of my marketers wants to launch a new marketing campaign or a website or whatever, they don’t wait around for an engineer to build an app or a website for them, my marketer builds it themselves.
My CFO no longer, you know, spends hours clicking through documents to do routine processing, and nor does she have to go around to evaluate lots of vendors to see who needs to pay tons of money to for some little automation.
She and her team, they write code themselves to automate financial processing. So I’m already seeing very clearly in Silicon Valley, my business and many other businesses, a noticeable and growing productivity gap between people, not software engineers, but people that know how to use AI to build custom software for them, and people that don’t.
And maybe if I’m wrong in two years, you can all come and yell at me, but I believe this will spread, and so I think there’s an imperative to basically teach everyone to learn to code. Can I add to that? Yes, so that’s an important one.
Andrew Ng saying that everyone needs to learn to code. That’s his big prescription here, Davos 2026. Yes, Himanshu, go ahead.
The other adjustment, a lot of the computer science graduates who are graduating, graduate immediately thinking they’re gonna get a job at a Meta or an Amazon or a Google or some company like that. Those jobs are gone. And then there’s a lot of disappointment where they wait for months, and in some cases years, waiting for those jobs to come back.
At the same time, there are HR departments, finance departments, sales department looking for AI skillset, and these people could get a hero status within that. We face that challenge today. You know, when we ask our CHRO that we don’t want recruiters, we want all the recruiting to be done using agents and bots, but someone has to code that.
And when they go out and look for people, they don’t find those computer science graduates because they’re waiting for their next big job to come from the big tech box. So that’s the reset I’m talking about, that once that starts happening, I think you’ll see a more even playing field in the world of skills and jobs.
Yeah, so don’t wait for the big tech job. Go outside of that universe as well. So I think there is a mentality and a mindset reset that’s also required for those who are coming into the workforce.
I’m gonna throw this open to questions. I can see a hand and many hands already being raised. Wonderful.
Can we get a microphone across to the lady, please? Thank you so much.
Hello, I am Derby Chukwudi, one of the Global Shapers that are present in Davos this week. And I’m based in Hong Kong. And one question I have is, if everybody learns how to code, and I’m just thinking from an economic standpoint, where the whole concept of comparative advantage and how we able to still showcase or focus on areas where we’re either very skilled or have leverage versus everyone trying to be the coder.
And then the second piece of the question is, what does the future of work look like if the typical jobs like the big tech jobs are gone or entry-level jobs as you think about it are no longer? What is the future? How should young people start thinking about things?
There’s a concept of the new economy, creator’s economy, but is that sustainable enough to build a career that can span through different phases of industries?
Thank you. Andrew, would you like to start?
Sure, to comparative advantage, coding skills goes really deep. My most productive developers, they’re actually not fresh college grads. They have 10, 20 years of experience in coding and are on top of AI.
It turns out there’s a one tier down from them is the fresh college grads that really know how to use AI. And then I’ll tell you one tier down from that is the people with 10 years of experience, but maybe they have a comfortable job and have not learned to embrace AI. So actually quite a few times my teams have voted to hire a fresh college grad that really knows AI over like a full stack developer or something that for some reason is still coding like it’s 2022 before modern AI.
And then unfortunately, the least productive that I would never hire are the fresh college grads that also do not know AI. So I feel like my market does not code nearly as well as a really good developers that I know with 10, 20 years of experience. So there is that deep expertise experience that still makes a huge difference.
Let me give a personal example where you sort of debunk this myth on what everyone’s accorded really means. So I have a son who graduated with a degree in cognitive science and he’s now working in a research lab building behavioral models, but he uses Python for that. He never learned coding.
And with a father who sort of professes on proper education and proper coding style, he uses chat GP to generate code. He examines the code and then he wants it. So is he a computer science engineer?
Is he a software coder or is he a cognitive scientist? I think with the concept of everyone codes, it doesn’t mean you’re writing necessarily complex algorithmic structured code. Now you have the ability to get started, jump in and write code in any domain.
And I think that sea change is gonna start impacting jobs. And when people start accepting that, then you could be in any profession being able to contribute to coding without being a traditional coder.
So dads are under threat from chat GPT as well. Yeah, so go ahead. Do you have a question?
Yeah, to the lady here and then I’ll get to you. So go ahead.
Thank you for sharing the insights and the information. I’m the founder of Waffle, which is the two-crucial gender gap industry in Japan. And also we engage in policy recommendation in Japan.
So my question is, what do you believe is the appropriate level and depth of AI education required at each educational stage, elementary school and junior high and senior high and school and university?
And the one more question is, how should a systematic continuity and progression in AI education be structured from elementary school through university? And I wanna ask Sarah and Andrews.
Please, Sarah, would you like to start? Yeah, sure.
So I’ll give you the basic outcomes that we’ve tackled. Do I have an answer of how it needs to evolve? No, our AI curriculum is a living curriculum because of that, because it needs to transform as we’re moving forward.
But I’ll take you stage by stage. If you’re talking about kindergarten students, it’s very important for them to understand, and we don’t call it AI for children, we call it robots. That there are robots that do things and the robots that do things need adult supervision for them to be able to interact with it.
As they’re moving older, especially for younger kids who are getting exposed, it’s very important for them to understand how does AI, in a very simple way, how does AI work, which take us back to machine learning.
So we teach them machine learning, which is AI builds a perspective based on the information that it’s been fed in. If it’s fed in single type of information, and we do that with the typical green apples, and is the red apple an apple or not. Then you move on to middle school and high school, and that’s where you start building in the critical thinking element.
Does it align with your ethics? Does it align with your values? Is the response you’ve gotten or the prompt that you’ve provided sufficient enough for it to give you the depth of knowledge that you require?
In what context are you supposed to use it? In what context is it really plagiarism and you’re not putting the effort in? Then comes in the rethinking of what an educator needs to be exposed to at different stages of their life to be able to tackle AI.
This is something that you’ve all mentioned, which is the expertise level. There needs to be a minimum level of knowledge required for people, especially today, to be able to use AI effectively, and be able to use it as an effective tool. So what we’re focusing on and doubling down on is students having a core body of knowledge, and then building on harnessing those skills.
I’ve mentioned critical thinking as one, resilience as another, and adaptability. And resilience and adaptability and critical thinking together actually builds the ability to learn and relearn within students. And that’s more of a long-term transformation of an education system that needs to be built in that has so many different elements that will take time and pieces within the system to take us out of the box of what education looks like today into what it needs to look like to give us the outcomes that we’re looking for.
Yes, could we have the microphone? Can I, we have about six minutes left, so I’ll try and take some more questions and come back to Andrew if there’s time left. Thank you.
Thank you very much for sharing your perspectives. We do a lot of work as well on assessing AI’s impact on the workforce, and we do understand that there are some jobs that are likely to sit in the augmentation category, and others which are at the risk of being displaced.
My question, and came to get your perspectives on it, is when human capital policies typically were designed, it was designed technically for humans, but we are now in an era of AI, which basically means humans will interact with robots, agentic AI.
A broad question is how do you see the human capital policies evolve, be it on the recruitment, career progression, retention in an AI-enabled world?
Professor, do you want to address that? Yeah, I mean, obviously it would change because the labor market is restructuring the way you use skills are restructuring, but I mean, agentic AI is creating jobs so far. You know, it’s like having an assistant, and you become more productive, and you produce more in companies that bring you in first, expand and all that.
So I don’t really see any big reforms that you have to do now for the future. The importance are now in education, and I want to go back to how much do you teach of AI and programming. I agree entirely that everyone should learn programming skills in the way that every kid in England has to learn French.
Do you think they ever use it? Or would they ever do as good a translation as AI can do from English to French, or French to English? No, but they learn it.
They learn Latin. I’ve got a son in school in England. It’s compulsory to do Latin.
I let them do programming as well. He’s doing programming as well, actually, in my son’s school, but maybe it’s a more progressive school. So that when you go out and get into a job, when you come across people speaking foreign languages, you don’t panic the way that so many kids that come out panic.
That’s the point of teaching a foreign language, and when you see AI, they don’t panic, because they’ve got the programming, and they know more about it. So that’s where the reform should take place, combined with experience that I was mentioning before. Once they’re in a company, then their human capital will be augmented all the time, because they will be learning new skills as new technology is arriving.
But that’s always existed in history. You know, when people driving cars, or horses carrying things in the late 19th century were told there is this wonderful thing now that uses internal combustion and drives, and give up your horse and come and drive it, it was a much bigger shock to them than what we’re facing now.
And yet, within 10 years, the horses disappeared, and roads were congested with cars. Well, maybe 20 years, but yeah.
Can I add one point?
Yes, please.
A lot of the policies today are static and progressional. You hire an employee, an employee gets promoted. You have events that happen in a very sequential manner.
If you think about it from the beginning of time, output of work was always a human, times efficiency, times productivity. We spend 100 years improving efficiency and productivity. For the first time, we are changing human, and adding an agent to that.
So all these policies now need to be reexamined, because they’re not gonna move linearly. There’s gonna be a lot more disruption. And again, there were several speakers today who touched on that.
So we are a believer, again, at Cornerstone, that this people graph has to be very dynamic, and that’s putting a lot of pressure on organizations around the world.
Can I add to why I think policy is so important? About two weeks ago, my team, we do a lot of person-in-the-street interviews. We spoke of a coffee shop owner in part of Silicon Valley, Mount Evie, California, who was politically shaking, because he was so angry at AI, with my team being representatives, and he was almost yelling at us for AI destroying the livelihoods of his artist friends.
I know here at WEF, we’re all very AI-positive, pro-AI, we see the business value, optimism for the future. I think many people underestimate the degree to which our sector is mistrusted, and sometimes even hated. One good news is, I think Adelman ran a study showing that the more people learn about AI, the less they distrust it, and the more they like it.
And so to gain societal acceptance of AI, to let us unlock the things that actually move society forward, I think this educational piece and the policy piece of it is really important.
Yes, building on trust as well is important. Yes, I’m going to give you an opportunity to get the microphone, and that’ll be the last question that we take for this evening.
Thank you very much for your insights and learning. I’m Kian, the founder and CEO of Orkera. I think if we look at our skill sets, we’d see that part of our skills are durable, they will be useful in a long time.
And I agree with Professor Eng that coding is becoming a durable skill now. But other skills are perishable, and they change so fast with a half-life that is so low. So my questions for you all is, how do you personally keep up with the change and acquire those perishable skills that are going to go away even six months from now?
Himanshu, or Melissa, or, yes, Professor.
I would repeat what Professor said, so he doesn’t kick me in the backside. Is the art of learning is learning how you learn, and as long as you’re good at that, what you’re going to learn is going to change. That cone of uncertainty is so broad right now that not only do we not know what the next jobs are, we don’t know what outcomes our jobs are going to generate.
So it’s very hard to teach a student about that. It’s easier to teach a student how to learn what’s the area of discipline and all the soft skills that you heard.
Yeah, I mean, people specialise in different things. By now, I have the courage to tell anyone who’s going to ask me questions, don’t ask me anything about this. I love that.
I love that. I love that. I love that.
And if they do, I say, out. I love that. I love that.
Yes. I mean, you cannot venture an answer to everything and expect to be listened to and be taken seriously. You have to admit that there are some things you don’t know.
Yes. I run Deep Learner.ai, world’s leading AI training platform, the chairman of Coursera. So I actually eat a lot of our own dog food.
I find that taking our courses actually helps me keep up with the cutting edge of the latest developments.
All right. With that, ladies and gentlemen, we are going to have to close this session, but thank you to our panel for joining us here this evening. Thank you very much for your participation.
This is an ongoing debate. I don’t think we have the answers yet. There’s going to be many more questions and hopefully we’ll be back in Davos again to address some of those.
From all of us here, goodbye and many thanks for watching.
Andrew Ng
Speech speed
206 words per minute
Speech length
1831 words
Speech time
532 seconds
Entry-level jobs exist but there’s a skills gap – workers lack AI skills needed for modern positions
Explanation
Andrew Ng argues that entry-level positions are available, but many businesses cannot find skilled workers with the necessary AI capabilities. He believes higher education is failing graduates by preparing them for pre-AI jobs rather than current market demands.
Evidence
He states he won’t hire engineers who don’t know how to use AI for coding, and that AI tools have advanced rapidly in software engineering
Major discussion point
AI’s Impact on Entry-Level Jobs and Employment
Topics
Economic | Future of work
AI will augment rather than replace most jobs – people using AI will replace those who don’t
Explanation
Ng contends that for most jobs, AI can handle 30-40% of tasks, meaning humans are still needed for the remaining 60-70%. However, workers who know how to use AI will have a competitive advantage over those who don’t.
Evidence
Task-based analysis shows AI can do 30-40% of most jobs, leaving 60-70% for humans to handle
Major discussion point
AI’s Impact on Entry-Level Jobs and Employment
Topics
Economic | Future of work
Agreed with
– Christopher Pissarides
– Himanshu Palsule
Agreed on
Most jobs will be augmented rather than completely replaced by AI
Certain jobs face complete displacement – translators, voice actors, call center operators at risk
Explanation
While most jobs will see partial automation, Ng identifies specific roles where AI can perform nearly 100% of the tasks. He emphasizes society has a moral obligation to support these displaced workers.
Evidence
Specifically mentions translators, voice actors, and call center operators as jobs where AI can do almost 100% of the work
Major discussion point
AI’s Impact on Entry-Level Jobs and Employment
Topics
Economic | Future of work
Current job disruptions are more from pandemic over-hiring than AI displacement
Explanation
Ng suggests that recent layoffs in tech and other sectors are primarily due to companies over-hiring during the pandemic rather than AI-driven job displacement. This challenges the narrative that current job losses are primarily AI-related.
Evidence
References layoffs over the past year or two being attributed to pandemic over-hiring rather than AI effects
Major discussion point
AI’s Impact on Entry-Level Jobs and Employment
Topics
Economic | Future of work
Higher education is failing students by preparing them for pre-AI jobs rather than current market needs
Explanation
Ng criticizes universities for not updating their curricula to include AI skills, leaving graduates unprepared for the modern job market. He sees this as a critical failure that needs urgent attention.
Evidence
States that higher education is preparing students for jobs of 2022 before modern AI, rather than jobs of 2026 and beyond
Major discussion point
Education System Reform and AI Literacy
Topics
Sociocultural | Online education
Agreed with
– Sarah bint Yousif Al Amiri
– Christopher Pissarides
– Himanshu Palsule
Agreed on
Education systems must fundamentally reform to include AI literacy and skills
Everyone should learn to code as it’s becoming essential across all professions, not just software engineering
Explanation
Ng advocates for universal coding education, arguing that AI-assisted coding is becoming valuable across all business functions. He observes a growing productivity gap between employees who can code with AI assistance and those who cannot.
Evidence
Examples of marketers building their own websites, recruiters writing code to screen resumes, and CFOs automating financial processing rather than hiring vendors
Major discussion point
Education System Reform and AI Literacy
Topics
Sociocultural | Online education
Disagreed with
– Audience
– Christopher Pissarides
Disagreed on
Universal coding education necessity and economic implications
Companies need to move beyond bottom-up point solutions to top-down business process redesign for transformative results
Explanation
Ng argues that most AI implementations focus on optimizing individual tasks, yielding only 5-10% efficiency gains. True transformation requires redesigning entire workflows from scratch, which requires both bottom-up innovation and top-down strategic thinking.
Evidence
Example of loan underwriting process where automating one step provides cost savings, but redesigning the entire process can offer 10-minute loan approval instead of one week, creating a new product
Major discussion point
Business Transformation and AI Implementation
Topics
Economic | Digital business models
Society has moral obligation to support workers whose jobs are completely displaced by AI
Explanation
Ng emphasizes that when AI can perform nearly 100% of certain jobs, society must take responsibility for supporting displaced workers. He argues that job displacement doesn’t mean workers deserve to be abandoned.
Evidence
Mentions specific concern for translators, voice actors, and call center operators whose jobs face complete automation
Major discussion point
Societal and Policy Implications
Topics
Economic | Future of work
Agreed with
– Sarah bint Yousif Al Amiri
Agreed on
Society has moral obligation to support displaced workers
AI sector faces significant public mistrust and sometimes hatred – education can help build acceptance
Explanation
Ng highlights the disconnect between the AI-positive atmosphere at forums like Davos and public sentiment, where many people distrust or even hate AI. He believes education can help bridge this gap and build societal acceptance.
Evidence
Personal anecdote of a coffee shop owner in Silicon Valley who was angry at AI for destroying his artist friends’ livelihoods; references Adelman study showing people like AI more as they learn about it
Major discussion point
Societal and Policy Implications
Topics
Sociocultural | Content policy
Christopher Pissarides
Speech speed
153 words per minute
Speech length
1638 words
Speech time
640 seconds
Most jobs involve tasks AI cannot fully automate – healthcare, hospitality, retail require human empathy and customer interaction
Explanation
Pissarides argues that approximately half the workforce performs jobs that don’t require AI and won’t be threatened by it. These jobs in healthcare, hospitality, and retail require human qualities like empathy and customer understanding that AI cannot replicate.
Evidence
Examples of nurses needing empathy to understand patients, hospitality workers needing to treat customers well to attract business, emphasizing AI has no evidence of having feelings
Major discussion point
AI’s Impact on Entry-Level Jobs and Employment
Topics
Economic | Future of work
Agreed with
– Andrew Ng
– Himanshu Palsule
Agreed on
Most jobs will be augmented rather than completely replaced by AI
Disagreed with
– Himanshu Palsule
Disagreed on
Scope of jobs that will be disrupted by AI
Professional services face the greatest disruption – law and accounting structured on outdated 19th century models
Explanation
Pissarides identifies traditional professions like law and accounting as most vulnerable to AI disruption because they were structured according to 19th century British educational models. He suggests these professions need fundamental reform.
Evidence
Describes the historical structure of British professions based on private school, Oxford/Cambridge classics education, leading to privileged positions in law and accounting
Major discussion point
AI’s Impact on Entry-Level Jobs and Employment
Topics
Economic | Future of work
Universities need reform to teach variety of skills combined with industry experience and internships
Explanation
Pissarides advocates for extending undergraduate education and incorporating more industry collaboration through internships and apprenticeships. This would help students gain both diverse skills and practical experience.
Evidence
Suggests extending British undergraduate education from three to four years, addressing the common complaint from graduates that employers want experience but won’t hire without it
Major discussion point
Education System Reform and AI Literacy
Topics
Sociocultural | Online education
Agreed with
– Sarah bint Yousif Al Amiri
– Himanshu Palsule
Agreed on
Continuous learning and adaptability are essential for future workforce
Disagreed with
– Andrew Ng
– Audience
Disagreed on
Universal coding education necessity and economic implications
Geopolitical fragmentation and deglobalization add complexity to AI workforce disruption
Explanation
Pissarides notes that companies now face the dual challenge of determining not just how to produce and what skills they need, but also where to produce and what supply chains to use. This geopolitical uncertainty complicates workforce planning beyond just AI considerations.
Evidence
References Macron’s talk about Europe needing deeper integration and single markets due to trust issues with behaviors ‘east and west’
Major discussion point
Societal and Policy Implications
Topics
Economic | Digital business models
Himanshu Palsule
Speech speed
174 words per minute
Speech length
1303 words
Speech time
447 seconds
Young people are most capable of implementing AI but are being left behind while companies focus on re-skilling management
Explanation
Palsule argues there’s an irony in that the generation most capable of adopting AI (young people) is being overlooked while companies spend resources re-skilling middle and senior management, which is more difficult. He advocates for focusing on young people’s natural AI capabilities.
Evidence
Suggests introducing junior workers to roles involving judgment and decision-making rather than autonomous repeatable roles that will disappear
Major discussion point
AI’s Impact on Entry-Level Jobs and Employment
Topics
Economic | Future of work
Agreed with
– Andrew Ng
– Sarah bint Yousif Al Amiri
– Christopher Pissarides
Agreed on
Education systems must fundamentally reform to include AI literacy and skills
Skills divide is widening due to unprecedented automation – need global approach to skills development
Explanation
Palsule emphasizes that skill shortages are real and driven by rapid automation, requiring a global perspective on skills development. He warns that protectionist policies could worsen the problem by limiting access to global talent.
Evidence
Examples of complex dialysis machines sitting idle due to lack of trained nurses, aircraft manufacturers lacking maintenance staff for new avionics
Major discussion point
Skills Gap and Workforce Development
Topics
Economic | Future of work
Agreed with
– Sarah bint Yousif Al Amiri
– Christopher Pissarides
Agreed on
Continuous learning and adaptability are essential for future workforce
Computer science graduates are waiting for big tech jobs that no longer exist while other departments need AI skills
Explanation
Palsule identifies a mismatch where computer science graduates wait for traditional big tech positions that have disappeared, while HR, finance, and sales departments desperately need AI skills. This represents a mindset problem that needs addressing.
Evidence
Personal experience at his company where they want AI-powered recruiting but can’t find computer science graduates because they’re waiting for jobs at Meta, Amazon, or Google
Major discussion point
Skills Gap and Workforce Development
Topics
Economic | Future of work
Focus should be on judgment, decision-making, and context skills rather than competing with AI on repetitive tasks
Explanation
Palsule advocates for developing skills that complement rather than compete with AI. He emphasizes that jobs involving words, language, and numbers will be disrupted, so humans should focus on areas where they add unique value.
Evidence
References Yuval Harari’s point that any job involving words, language, and numbers will be disrupted; suggests having junior analysts examine AI output rather than compete with AI
Major discussion point
Skills Gap and Workforce Development
Topics
Economic | Future of work
Agreed with
– Andrew Ng
– Christopher Pissarides
Agreed on
Most jobs will be augmented rather than completely replaced by AI
Disagreed with
– Christopher Pissarides
Disagreed on
Scope of jobs that will be disrupted by AI
Human capital policies need fundamental restructuring as work output now includes human plus AI agent collaboration
Explanation
Palsule argues that traditional HR policies designed for humans alone are inadequate for the new reality where work output involves human-AI collaboration. This requires dynamic rather than static, sequential approaches to career progression.
Evidence
Notes that for the first time in history, the work output formula is changing from ‘human times efficiency times productivity’ to include AI agents
Major discussion point
Business Transformation and AI Implementation
Topics
Economic | Future of work
Need for cross-border skills mobility as talent gaps exist globally while policies become more protectionist
Explanation
Palsule highlights the contradiction between global skills shortages and increasingly protectionist policies that limit talent mobility. He advocates for thinking about skills development and deployment on a global scale.
Evidence
Examples of skilled analysts in Nigeria or Riyadh who might have better skills than locally hired people; mentions his company hiring globally with half their teams outside the US
Major discussion point
Societal and Policy Implications
Topics
Economic | Future of work
Sarah bint Yousif Al Amiri
Speech speed
181 words per minute
Speech length
1220 words
Speech time
402 seconds
UAE is first country to mandate K-12 AI literacy for all students, teaching proper AI usage and ethics
Explanation
Al Amiri describes the UAE’s pioneering approach to AI education, making it mandatory for all students from kindergarten through 12th grade. The program focuses on teaching students how to use AI tools effectively and ethically.
Evidence
280,000 students taking AI literacy classes at least once every two weeks; curriculum covers how to use AI, when to use it, and what tools to use
Major discussion point
Education System Reform and AI Literacy
Topics
Sociocultural | Online education
Agreed with
– Andrew Ng
– Christopher Pissarides
– Himanshu Palsule
Agreed on
Education systems must fundamentally reform to include AI literacy and skills
AI education should include critical thinking skills to evaluate AI outputs and avoid bias
Explanation
Al Amiri emphasizes that AI education must go beyond technical skills to include critical thinking about AI outputs. Students learn to evaluate different prompts, identify bias, and understand the quality of AI-generated information.
Evidence
Example of students using different prompts about World War I and discussing with teachers which prompts worked better and what information was biased or accurate
Major discussion point
Education System Reform and AI Literacy
Topics
Sociocultural | Online education
Education must focus on teaching students how to learn, relearn, and unlearn continuously
Explanation
Al Amiri argues that the education system must shift from knowledge transmission to developing meta-learning skills. Students need to be able to adapt continuously rather than rely on static knowledge acquisition.
Evidence
Describes the shift from asking students to repeat information to helping them pick out relevant information, understand bias, and analyze properly
Major discussion point
Education System Reform and AI Literacy
Topics
Sociocultural | Online education
Agreed with
– Christopher Pissarides
– Himanshu Palsule
Agreed on
Continuous learning and adaptability are essential for future workforce
Government has obligation to reskill workers and foot the bill for retraining programs
Explanation
Al Amiri acknowledges that governments have a responsibility to provide retraining for workers displaced by technological change. However, she emphasizes the importance of building intrinsic learning capabilities to reduce future government intervention costs.
Evidence
References UAE program launched after COVID to reskill individuals for private sector jobs, with government funding the retraining
Major discussion point
Skills Gap and Workforce Development
Topics
Economic | Future of work
Agreed with
– Andrew Ng
Agreed on
Society has moral obligation to support displaced workers
Shereen Bhan
Speech speed
187 words per minute
Speech length
1579 words
Speech time
505 seconds
There is palpable anxiety about AI’s impact on jobs market across near, medium, and long term
Explanation
Bhan highlights the widespread concern and uncertainty about how AI will affect employment at different time horizons. She notes this anxiety is particularly prominent in discussions at Davos 2026.
Evidence
References the World Economic Forum’s jobs report and the general atmosphere of concern at Davos
Major discussion point
AI’s Impact on Entry-Level Jobs and Employment
Topics
Economic | Future of work
There’s an irony that while AI threatens entry-level jobs, significant skill gaps exist in sectors like healthcare
Explanation
Bhan points out the paradox that while there are fears about AI eliminating jobs, there are simultaneously massive shortages of skilled workers in critical sectors. This suggests the problem is more complex than simple job displacement.
Evidence
WEF report showing an 11 million worker gap in healthcare sector alone
Major discussion point
Skills Gap and Workforce Development
Topics
Economic | Future of work
Current discussions about AI transformation are still focused on efficiency and optimization rather than true transformation
Explanation
Bhan observes that despite talk of transformation, most conversations with business leaders still center on bottom-up improvements and efficiency gains rather than fundamental business model changes. She questions whether organizations have enough time to prepare for more transformative impacts.
Evidence
Her ongoing conversations with business leaders showing focus on efficiency rather than transformation
Major discussion point
Business Transformation and AI Implementation
Topics
Economic | Digital business models
Geopolitical risks and economic fragmentation compound AI’s impact on labor markets
Explanation
Bhan argues that AI’s impact on jobs cannot be viewed in isolation but must be considered alongside broader economic disruptions including geopolitical tensions, geoeconomic confrontation, and retreat from globalization. These combined factors create a structurally disruptive environment for labor markets.
Evidence
References to geoeconomic fragmentation and retreat from globalization happening simultaneously with AI adoption
Major discussion point
Societal and Policy Implications
Topics
Economic | Future of work
Audience
Speech speed
175 words per minute
Speech length
474 words
Speech time
161 seconds
Universal coding education may undermine comparative advantage and economic specialization
Explanation
An audience member questions whether everyone learning to code conflicts with economic principles of comparative advantage, where individuals and societies benefit from specializing in areas where they have relative strengths. They worry about the sustainability of everyone trying to become coders.
Evidence
References economic concept of comparative advantage and questions about career sustainability in creator economy
Major discussion point
Education System Reform and AI Literacy
Topics
Economic | Future of work
Disagreed with
– Andrew Ng
– Christopher Pissarides
Disagreed on
Universal coding education necessity and economic implications
Need for systematic progression in AI education from elementary through university levels
Explanation
An audience member from Japan asks about the appropriate depth and continuity of AI education across different educational stages. They seek guidance on how to structure AI learning progression from elementary school through university to ensure systematic skill development.
Evidence
Specific question about educational stages: elementary, junior high, senior high, and university
Major discussion point
Education System Reform and AI Literacy
Topics
Sociocultural | Online education
Human capital policies need redesign for AI-human collaboration in recruitment, career progression, and retention
Explanation
An audience member highlights that traditional HR policies were designed for human-only workforces but now need to account for human-AI collaboration. They question how policies around recruitment, career advancement, and employee retention should evolve in an AI-enabled workplace.
Evidence
Distinction between jobs at risk of displacement versus those in augmentation category
Major discussion point
Business Transformation and AI Implementation
Topics
Economic | Future of work
Challenge of maintaining durable skills while continuously acquiring perishable skills with short half-lives
Explanation
An audience member distinguishes between durable skills that remain useful long-term and perishable skills that become obsolete quickly in the AI era. They seek advice on how professionals can balance investing in lasting capabilities while continuously updating rapidly-changing technical skills.
Evidence
Recognition that some skills have very short half-lives, becoming obsolete within six months
Major discussion point
Skills Gap and Workforce Development
Topics
Economic | Future of work
Agreements
Agreement points
Education systems must fundamentally reform to include AI literacy and skills
Speakers
– Andrew Ng
– Sarah bint Yousif Al Amiri
– Christopher Pissarides
– Himanshu Palsule
Arguments
Higher education is failing students by preparing them for pre-AI jobs rather than current market needs
UAE is first country to mandate K-12 AI literacy for all students, teaching proper AI usage and ethics
Universities need reform to teach variety of skills combined with industry experience and internships
Young people are most capable of implementing AI but are being left behind while companies focus on re-skilling management
Summary
All speakers agree that current education systems are inadequate for the AI era and require comprehensive reform from K-12 through university levels to prepare students with relevant AI skills and capabilities
Topics
Sociocultural | Online education
Most jobs will be augmented rather than completely replaced by AI
Speakers
– Andrew Ng
– Christopher Pissarides
– Himanshu Palsule
Arguments
AI will augment rather than replace most jobs – people using AI will replace those who don’t
Most jobs involve tasks AI cannot fully automate – healthcare, hospitality, retail require human empathy and customer interaction
Focus should be on judgment, decision-making, and context skills rather than competing with AI on repetitive tasks
Summary
Speakers consensus that while AI will significantly impact work, most jobs will involve human-AI collaboration rather than complete human replacement, with humans focusing on tasks requiring judgment, empathy, and contextual understanding
Topics
Economic | Future of work
Continuous learning and adaptability are essential for future workforce
Speakers
– Sarah bint Yousif Al Amiri
– Christopher Pissarides
– Himanshu Palsule
Arguments
Education must focus on teaching students how to learn, relearn, and unlearn continuously
Universities need reform to teach variety of skills combined with industry experience and internships
Skills divide is widening due to unprecedented automation – need global approach to skills development
Summary
All speakers emphasize that the ability to continuously learn and adapt is more important than acquiring static knowledge, as the pace of technological change requires ongoing skill development throughout careers
Topics
Economic | Future of work
Society has moral obligation to support displaced workers
Speakers
– Andrew Ng
– Sarah bint Yousif Al Amiri
Arguments
Society has moral obligation to support workers whose jobs are completely displaced by AI
Government has obligation to reskill workers and foot the bill for retraining programs
Summary
Both speakers agree that when AI displaces workers, there is a collective responsibility to provide support and retraining rather than abandoning affected individuals
Topics
Economic | Future of work
Similar viewpoints
Both speakers see coding as a universal skill that extends beyond traditional software engineering roles, with AI-assisted coding becoming valuable across all business functions and departments
Speakers
– Andrew Ng
– Himanshu Palsule
Arguments
Everyone should learn to code as it’s becoming essential across all professions, not just software engineering
Computer science graduates are waiting for big tech jobs that no longer exist while other departments need AI skills
Topics
Sociocultural | Online education
Both speakers recognize that traditional professional structures and policies are outdated and require fundamental restructuring to accommodate AI integration and modern work realities
Speakers
– Christopher Pissarides
– Himanshu Palsule
Arguments
Professional services face the greatest disruption – law and accounting structured on outdated 19th century models
Human capital policies need fundamental restructuring as work output now includes human plus AI agent collaboration
Topics
Economic | Future of work
Both speakers challenge the narrative of widespread AI-driven job displacement, arguing that current disruptions have other causes and that many jobs remain fundamentally human-centered
Speakers
– Andrew Ng
– Christopher Pissarides
Arguments
Current job disruptions are more from pandemic over-hiring than AI displacement
Most jobs involve tasks AI cannot fully automate – healthcare, hospitality, retail require human empathy and customer interaction
Topics
Economic | Future of work
Unexpected consensus
Universal coding education across all professions
Speakers
– Andrew Ng
– Christopher Pissarides
– Himanshu Palsule
Arguments
Everyone should learn to code as it’s becoming essential across all professions, not just software engineering
Universities need reform to teach variety of skills combined with industry experience and internships
Computer science graduates are waiting for big tech jobs that no longer exist while other departments need AI skills
Explanation
Unexpected that an economist (Pissarides) would strongly support universal programming education, comparing it to mandatory foreign language learning. This consensus across technical and economic perspectives suggests coding is viewed as a fundamental literacy rather than specialized skill
Topics
Sociocultural | Online education
AI hype exceeds current reality of job displacement
Speakers
– Andrew Ng
– Christopher Pissarides
Arguments
Current job disruptions are more from pandemic over-hiring than AI displacement
Most jobs involve tasks AI cannot fully automate – healthcare, hospitality, retail require human empathy and customer interaction
Explanation
Surprising consensus between AI expert and economist that current AI job displacement fears are overblown, with both attributing recent layoffs to other factors and emphasizing human-centric aspects of work that remain irreplaceable
Topics
Economic | Future of work
Overall assessment
Summary
Strong consensus on education reform needs, human-AI collaboration model, and continuous learning importance. Moderate agreement on societal obligations and skills development approaches.
Consensus level
High level of consensus among speakers suggests a mature understanding of AI’s impact on work, moving beyond apocalyptic predictions to practical solutions focused on adaptation, education reform, and human-AI collaboration. This alignment across diverse expertise areas (AI technology, economics, government policy, business) indicates potential for coordinated policy responses and implementation strategies.
Differences
Different viewpoints
Scope of jobs that will be disrupted by AI
Speakers
– Christopher Pissarides
– Himanshu Palsule
Arguments
Most jobs involve tasks AI cannot fully automate – healthcare, hospitality, retail require human empathy and customer interaction
Focus should be on judgment, decision-making, and context skills rather than competing with AI on repetitive tasks
Summary
Pissarides argues that about half the workforce (healthcare, hospitality, retail) won’t be threatened by AI because these jobs require human qualities like empathy that AI cannot replicate. Palsule, referencing Yuval Harari, contends that any job involving words, language, and numbers will get disrupted, suggesting a much broader scope of AI impact across sectors including healthcare.
Topics
Economic | Future of work
Universal coding education necessity and economic implications
Speakers
– Andrew Ng
– Audience
– Christopher Pissarides
Arguments
Everyone should learn to code as it’s becoming essential across all professions, not just software engineering
Universal coding education may undermine comparative advantage and economic specialization
Universities need reform to teach variety of skills combined with industry experience and internships
Summary
Ng strongly advocates that everyone should learn coding as he sees a growing productivity gap between those who can use AI to code and those who cannot. An audience member challenges this by questioning whether universal coding conflicts with economic principles of comparative advantage. Pissarides takes a more moderate stance, comparing coding education to learning foreign languages – useful for familiarity but not expecting everyone to become expert coders.
Topics
Sociocultural | Online education
Unexpected differences
Primary cause of recent job layoffs
Speakers
– Andrew Ng
– Shereen Bhan
Arguments
Current job disruptions are more from pandemic over-hiring than AI displacement
There is palpable anxiety about AI’s impact on jobs market across near, medium, and long term
Explanation
This disagreement is unexpected because it challenges the fundamental premise of the discussion. While the entire panel is convened to discuss AI’s impact on jobs, Ng argues that current layoffs are primarily due to pandemic over-hiring rather than AI displacement. This contradicts the general anxiety about AI job displacement that Bhan identifies as driving the discussion, suggesting that current fears may be misattributed.
Topics
Economic | Future of work
Overall assessment
Summary
The main areas of disagreement center on the scope of AI’s job impact, the necessity of universal coding education, and the primary causes of current job disruptions. Speakers also differ on educational reform approaches and which worker demographics should be prioritized.
Disagreement level
Moderate disagreement with significant implications. While speakers generally agree on the need for education reform and that AI will augment rather than replace most jobs, their different views on scope and approach could lead to very different policy recommendations. The disagreement on whether current layoffs are AI-related versus pandemic-related is particularly significant as it questions the urgency of the problem being discussed.
Partial agreements
Partial agreements
Similar viewpoints
Both speakers see coding as a universal skill that extends beyond traditional software engineering roles, with AI-assisted coding becoming valuable across all business functions and departments
Speakers
– Andrew Ng
– Himanshu Palsule
Arguments
Everyone should learn to code as it’s becoming essential across all professions, not just software engineering
Computer science graduates are waiting for big tech jobs that no longer exist while other departments need AI skills
Topics
Sociocultural | Online education
Both speakers recognize that traditional professional structures and policies are outdated and require fundamental restructuring to accommodate AI integration and modern work realities
Speakers
– Christopher Pissarides
– Himanshu Palsule
Arguments
Professional services face the greatest disruption – law and accounting structured on outdated 19th century models
Human capital policies need fundamental restructuring as work output now includes human plus AI agent collaboration
Topics
Economic | Future of work
Both speakers challenge the narrative of widespread AI-driven job displacement, arguing that current disruptions have other causes and that many jobs remain fundamentally human-centered
Speakers
– Andrew Ng
– Christopher Pissarides
Arguments
Current job disruptions are more from pandemic over-hiring than AI displacement
Most jobs involve tasks AI cannot fully automate – healthcare, hospitality, retail require human empathy and customer interaction
Topics
Economic | Future of work
Takeaways
Key takeaways
AI will augment rather than replace most jobs – people who use AI will replace those who don’t, but AI itself won’t replace humans in most roles
Entry-level jobs exist but there’s a critical skills gap – workers lack the AI literacy needed for modern positions
Education systems are failing by preparing students for pre-AI jobs rather than current market needs
Everyone should learn to code as it’s becoming essential across all professions, not just software engineering
Certain jobs face complete displacement (translators, voice actors, call center operators) while others require human skills AI cannot replicate (healthcare, hospitality)
Companies need to move beyond incremental efficiency gains to transformative business process redesign using AI
Continuous learning and the ability to ‘learn how to learn’ is becoming the most critical skill for the future workforce
The UAE model of mandating K-12 AI literacy with critical thinking components should be adopted globally
Society has a moral obligation to support workers whose jobs are completely displaced by AI
Current job disruptions are more from pandemic over-hiring than actual AI displacement
Resolutions and action items
Governments should mandate AI literacy in K-12 education following the UAE model
Universities need to reform curricula to include AI skills, industry experience, and internships
Companies should create apprenticeships and environments for AI-fluent young workers rather than focusing only on re-skilling management
Businesses should invest in top-down workflow redesign rather than just bottom-up point solutions
Educational institutions should teach students how to learn continuously rather than just static knowledge
Governments should fund retraining programs for displaced workers
Companies should hire based on AI skills rather than waiting for traditional big tech positions
Unresolved issues
How to structure systematic AI education progression from elementary through university levels
What specific human capital policies should be redesigned for human-AI collaboration
How to address the growing skills divide in a world with increasing protectionist policies limiting talent mobility
How to gain broader societal acceptance of AI given current public mistrust and fear
What the sustainable career paths look like in the creator economy and new economic models
How to balance comparative advantage when everyone learns coding skills
How to keep up with rapidly changing ‘perishable skills’ that become obsolete within months
How to manage the transition period while businesses figure out transformative AI implementation
Suggested compromises
Focus on T-shaped skills model – broad AI literacy combined with deep domain expertise
Teach coding as a basic literacy skill (like foreign languages) rather than expecting everyone to become professional programmers
Combine university education with industry partnerships for practical experience
Have junior workers examine AI outputs rather than compete with AI on repetitive tasks
Extend university programs to include more practical training and internships
Create dynamic rather than static career progression policies that can adapt to AI disruption
Balance bottom-up innovation with top-down strategic AI implementation in businesses
Thought provoking comments
I’m not going to ever hire another engineer again that doesn’t know how to use AI to help them code… Unfortunately, I feel like higher education, which I love and I think is a great force for the good, is failing many fresh college graduates by preparing them for the jobs of 2022 before modern AI, rather than the jobs of 2026 and beyond.
Speaker
Andrew Ng
Reason
This comment reframes the entire discussion from ‘AI will eliminate jobs’ to ‘AI illiteracy will eliminate job prospects.’ It’s provocative because it directly challenges higher education institutions and suggests the problem isn’t AI displacement but educational lag.
Impact
This comment set the foundational tone for the entire discussion, shifting focus from fear-based narratives about job loss to skills-based solutions. It prompted the Minister to elaborate on UAE’s K-12 AI curriculum and established the central theme that education reform, not job protection, is the key issue.
We looked at social media. Students have used social media extensively, haven’t been given a code of ethics on how to use it effectively… we’ve seen adverse effects on both their personalities, their social well-being, and their social interactions as well. We saw that that might happen in their cognitive ability, the development of their cognitive ability when it comes to AI.
Speaker
Sarah bint Yousif Al Amiri
Reason
This analogy between social media and AI is deeply insightful because it draws from recent historical experience to predict future challenges. It suggests that unguided technology adoption can have profound developmental consequences, making the case for proactive education rather than reactive regulation.
Impact
This comment elevated the discussion from technical skills training to broader questions of cognitive development and ethical technology use. It provided a compelling rationale for why AI education should start early and be comprehensive, influencing subsequent discussions about curriculum design across educational stages.
I think the greatest irony of our time is we are leaving behind a generation that is actually the most capable of implementing AI and we are spending time re-skilling middle management and senior management, which is harder.
Speaker
Himanshu Palsule
Reason
This observation is counterintuitive and challenges conventional wisdom about corporate training priorities. It highlights a strategic misallocation of resources and suggests that organizations are approaching AI adoption backwards.
Impact
This comment introduced a critical perspective on corporate AI strategy and prompted discussion about how hiring and training practices need fundamental restructuring. It shifted the conversation from individual skill development to organizational strategy and resource allocation.
It’s time to get everyone to learn to code… we’re already seeing in Silicon Valley that not just the software engineers, but the marketers, HR professionals, financial analysts, and so on, the ones that know how to code are much more productive than the ones that don’t, and that gap is growing.
Speaker
Andrew Ng
Reason
This is perhaps the most controversial and transformative suggestion in the discussion. It challenges traditional professional boundaries and suggests coding is becoming as fundamental as literacy, which would represent a massive shift in educational and professional requirements.
Impact
This comment sparked immediate debate and became a central talking point for the remainder of the discussion. It prompted questions about comparative advantage, led to practical examples from other panelists, and forced the group to grapple with what ‘universal coding’ would actually mean in practice.
If you look at law, accounting… those professions were structured more or less by the British in the 19th century mainly and they did it according to the education that existed at the time… That’s the one that is hit most because that’s what generative AI and the large language models can do and in fact, it’s about time they got a kick in the backside and they reformed the professions.
Speaker
Christopher Pissarides
Reason
This historical perspective is brilliantly insightful because it contextualizes current disruption within centuries of professional evolution. The provocative language about professions needing ‘a kick in the backside’ challenges entrenched professional structures and suggests AI disruption might be beneficial for overdue reforms.
Impact
This comment brought historical depth to the discussion and provided a framework for understanding why certain sectors are more vulnerable to AI disruption. It shifted the conversation from defensive positioning to considering AI as a catalyst for necessary structural reforms in professional services.
About two weeks ago, my team… spoke of a coffee shop owner in part of Silicon Valley… who was politically shaking, because he was so angry at AI… I think many people underestimate the degree to which our sector is mistrusted, and sometimes even hated.
Speaker
Andrew Ng
Reason
This personal anecdote powerfully illustrates the disconnect between the optimistic AI discourse at elite forums like Davos and public sentiment. It introduces the critical dimension of social acceptance and trust, which could ultimately determine AI adoption success regardless of technical capabilities.
Impact
This comment brought a sobering reality check to the discussion and introduced the crucial topic of public trust and social acceptance. It shifted the conversation from purely technical and economic considerations to include social and political dimensions, emphasizing that education serves not just skill development but also trust-building functions.
Overall assessment
These key comments fundamentally shaped the discussion by transforming it from a typical ‘AI job displacement’ conversation into a nuanced exploration of educational reform, organizational strategy, and social acceptance. Andrew Ng’s opening comment about hiring practices set a solutions-oriented tone that prevented the discussion from becoming defensive or fear-based. The Minister’s social media analogy provided historical context that legitimized proactive educational intervention. Himanshu’s observation about generational capabilities challenged corporate assumptions, while the Professor’s historical perspective on professional structures reframed disruption as potentially beneficial reform. The coding universality proposal became a central debate point that forced concrete discussion about future skill requirements. Finally, the coffee shop anecdote grounded the elite discussion in social reality. Together, these comments created a multi-dimensional conversation that addressed technical, educational, organizational, historical, and social aspects of AI’s workforce impact, making it far more comprehensive and actionable than typical discussions on this topic.
Follow-up questions
How can we measure the effectiveness of K-12 AI literacy programs and their long-term impact on workforce readiness?
Speaker
Sarah bint Yousif Al Amiri
Explanation
The UAE Minister mentioned they have a ‘living curriculum’ that needs to transform as they move forward, indicating ongoing evaluation and adaptation needs for their pioneering K-12 AI program
What specific metrics and frameworks should be used to identify which AI projects can drive strategic value versus just cost savings?
Speaker
Andrew Ng
Explanation
Andrew mentioned the difficulty of sorting through hundreds of project ideas to identify those with strategic value, highlighting the need for better evaluation frameworks
How can we develop global standards for cross-border skills mobility and recognition in an AI-driven economy?
Speaker
Himanshu Palsule
Explanation
He emphasized the need to think about skills sovereignty similar to data sovereignty, suggesting research into frameworks for global skills mobility and recognition
What are the most effective models for university-industry collaboration in developing AI-relevant curricula and providing work experience?
Speaker
Christopher Pissarides
Explanation
He suggested extending university education and combining it with industry internships/apprenticeships, but didn’t specify the optimal models for such collaboration
How do we systematically identify and support workers in jobs where AI can perform 80-100% of tasks?
Speaker
Andrew Ng
Explanation
He specifically mentioned translators, voice actors, and call center operators as at-risk groups requiring societal support, but didn’t detail identification or support mechanisms
What is the optimal depth and progression of AI education from elementary through university levels?
Speaker
Audience member from Japan
Explanation
This question was raised but not fully answered, indicating need for research into systematic AI education progression across all educational stages
How should human capital policies evolve for recruitment, career progression, and retention in an AI-enabled workplace?
Speaker
Audience member
Explanation
The question addressed the need to redesign HR policies for human-AI collaboration, but received only partial responses requiring further investigation
What are the most effective strategies for individuals to continuously acquire perishable skills with short half-lives?
Speaker
Kian (Audience member)
Explanation
While panelists emphasized learning how to learn, specific methodologies and frameworks for rapid skill acquisition in fast-changing environments need further research
How can we build societal trust and acceptance of AI to enable broader adoption and reduce resistance?
Speaker
Andrew Ng
Explanation
He mentioned significant mistrust and even hatred toward AI in some communities, referencing a study showing education reduces distrust, but more research is needed on trust-building strategies
What are the economic implications of everyone learning to code, and how does this affect comparative advantage principles?
Speaker
Derby Chukwudi (Audience member)
Explanation
This economic question about widespread coding skills and comparative advantage was raised but not fully explored, requiring further economic analysis
Disclaimer: This is not an official session record. DiploAI generates these resources from audiovisual recordings, and they are presented as-is, including potential errors. Due to logistical challenges, such as discrepancies in audio/video or transcripts, names may be misspelled. We strive for accuracy to the best of our ability.
Related event
World Economic Forum Annual Meeting 2026 at Davos
19 Jan 2026 08:00h - 23 Jan 2026 18:00h
Davos, Switzerland