00:00:04 Veena McCoole
Hello and welcome to the Human Interface, brought to you by the Oxford Internet Institute at the University of Oxford.
00:00:11 Veena McCoole
This is a podcast about how developments in technology impact our work and life, as told through the research and insights of a brilliant roster of experts and industry practitioners from the university and beyond.
00:00:23 Veena McCoole
I'm your host, Veena McCool.
00:00:26 Veena McCoole
AI is reshaping every sector of the economy, but what does this mean for today's workforce and the skills they need to bring into their careers?
00:00:33 Veena McCoole
In this episode, we explore what actually matters in today's job market, which AI skills matter, how to think about acquiring them, and where institutions and governments can lead the charge in shaping how AI fits in the future of work.
00:00:46 Veena McCoole
Dr.
00:00:46 Veena McCoole
Fabian Stefani is a lecturer and researcher at the Oxford Internet Institute.
00:00:51 Veena McCoole
and author of the forthcoming book, Skilled, which argues that today's biggest labor market challenges can be understood through the lens of skills, and that the skills gap has knock-on effects on wages, employment, and economic mobility.
00:01:05 Veena McCoole
Pratik Sibal specializes in digital transformation in the communication and information sector at UNESCO, and can offer an insight into what AI upskilling looks like inside public institutions.
00:01:16 Veena McCoole
Fabian and Pratik, welcome to the show.
00:01:18 Fabian Stephany
Thanks.
00:01:19 Fabian Stephany
Thanks for having me.
00:01:20 Prateek Sibal
Thanks.
00:01:20 Prateek Sibal
Thanks, Weena.
00:01:22 Veena McCoole
So to get started, we know AI is thought to lead to more productivity, but what does this actually mean for workers and what skills are they expected to bring into the workplaces of today?
00:01:32 Veena McCoole
Fabian, let's start with you.
00:01:34 Fabian Stephany
So first of all, it's important to say there's no one-size-fits-all approach to this.
00:01:38 Fabian Stephany
So it's hard to say like, okay, for all workers on the labor market, I give the same recommendation.
00:01:42 Fabian Stephany
But
00:01:43 Fabian Stephany
There are some aspects that emerge from our research recently here at the OII, and let me focus in the interest of time, focus on two of them.
00:01:51 Fabian Stephany
One is the skills centered around artificial intelligence.
00:01:55 Fabian Stephany
And when I say this, I want to explicitly stress this is not only technical skills.
00:02:00 Fabian Stephany
So I'm not necessarily talking about how to develop AI being a machine learning engineer, but when I say AI skills, I also mean skills as a person in sales, in marketing, in research, in admin that allow you to facilitate your work, whatever you're doing, and integrate technology into this.
00:02:18 Fabian Stephany
And the other stream, what we see, maybe surprisingly to people, but if you think about it, it's not such a big surprise, is so-called complementary skills.
00:02:25 Fabian Stephany
So that are skills where the technology is not great at compared to a human being, which would be, for example, team management or ethical judgment.
00:02:35 Fabian Stephany
And we see that the more AI is entering certain industries and occupations, the more we see that these skills are being not only more demanded, but also more rewarded.
00:02:44 Fabian Stephany
So those would be at least two groups of skills that I think are safe to recommend to
00:02:48 Fabian Stephany
I guess almost everybody.
00:02:50 Prateek Sibal
And I think even before we go into what is the impact on workers, the debate around productivity improvements with the use of AI is not settled.
00:02:58 Prateek Sibal
There's a lot of evidence which still needs to be developed to say actually there is a productivity improvement in the economy.
00:03:06 Prateek Sibal
And the questions that we need to ask is what amount of that
00:03:11 Prateek Sibal
improvement in productivity is also translating into improved wages for workers.
00:03:16 Prateek Sibal
So I think these are some broader questions that first we need to ask.
00:03:19 Prateek Sibal
And then specifically to the workers, I think there is a lot of discussion around cognitive atrophy, which we need to take care about because
00:03:29 Prateek Sibal
we are talking about medium to long-term well-being of workers.
00:03:33 Prateek Sibal
So there's a lot of push around efficiency and use AI tools for your workplace.
00:03:37 Prateek Sibal
But then if you are not thinking about worker well-being, this is not good for societies.
00:03:43 Prateek Sibal
We need to focus on that.
00:03:44 Prateek Sibal
The second thing is also, I feel that over a period of time, because everything is so much easier to do with AI, we are not thinking so much about what is priority.
00:03:55 Prateek Sibal
So that's this double whammy where
00:03:57 Prateek Sibal
workers are being asked to do more and more and people are saying, yes, but just use Gen AI and you should be able to do this in like 10 minutes.
00:04:04 Prateek Sibal
But then is it really important for the company, but also for the worker?
00:04:08 Prateek Sibal
These are some of the questions that we need to, I think, reflect more deeply upon when we are talking about workers.
00:04:14 Fabian Stephany
Yeah.
00:04:14 Fabian Stephany
And if I may just jump in on this, because it's really fascinating.
00:04:17 Fabian Stephany
There's something that comes up again and again in economic history, which is called so-called Jevons paradox, which essentially just says like,
00:04:25 Fabian Stephany
we have the means to be more efficient in doing things.
00:04:29 Fabian Stephany
And you could think that in the case of AI, for example, that would lead us to having more free time.
00:04:34 Fabian Stephany
But it seems to be currently that the opposite might be the case, right?
00:04:38 Fabian Stephany
It's like these tools are readily available and without guidance.
00:04:42 Fabian Stephany
And that's what I really like what you said, Pritik, without guidance, people just throw these tools at everything.
00:04:47 Fabian Stephany
And that might lead to the contrary, people feeling that they are busier than before because they now get things done
00:04:55 Fabian Stephany
probably are really faster and maybe at higher quality, but they're also, in certain circumstances, being demanded for just doing more.
00:05:01 Fabian Stephany
So it's not necessarily the case that AI is going to improve well-being in the sense that, oh, I'll get my things done quicker and at a similar or maybe higher quality and have more leisure time afterwards.
00:05:12 Prateek Sibal
Absolutely.
00:05:12 Prateek Sibal
And that's a bit in the history of automation.
00:05:14 Prateek Sibal
Like more and more we start using digital technologies, life becomes easier for sure, but then it is not freeing us time.
00:05:21 Prateek Sibal
So now you go to the supermarket and do your own checkout or you
00:05:25 Prateek Sibal
you do your own check-in for the flights or the hotels.
00:05:29 Prateek Sibal
This is more work, which is easier to do, so it gets piled on to you, which was supposed to be part of someone's job before.
00:05:36 Prateek Sibal
So I'm not saying this is bad or good, but I think we just need to be conscious of where we are heading and what do we want, and then decide consciously, yes, this is the future we are okay with.
00:05:47 Veena McCoole
Absolutely.
00:05:48 Veena McCoole
And there's been so much in the news recently about the impact of AI skills and what that has done for things like the graduate job market or how that has kind of concentrated the number of tasks that any given person does, as you said, Fabian, in their roles.
00:06:02 Veena McCoole
What does that mean for actual vacancies?
00:06:05 Veena McCoole
So I guess if I was a job seeker and I wanted to increase my chances of finding work, perhaps, what options do I have and where should I really be focusing?
00:06:14 Fabian Stephany
There are two things that are emerging right now in the research that we see around, particularly around entrance jobs, like entry jobs, people graduating and trying to find their first, footing their first job on the labor market.
00:06:25 Fabian Stephany
One is, and there's been a widely cited paper by colleagues at Stanford around Erik Brindelofsen where they looked at how entry-level jobs, the demand for entry-level jobs has been apparently declined.
00:06:37 Fabian Stephany
They tried to causally relate this to AI, which
00:06:40 Fabian Stephany
there's certain caveats around this study, but let's assume that AI plays a certain role in this, because there is an economic incentive for firms, particularly when it comes to entry-level jobs, to say, okay, we might want to automate away some of these jobs, because the tasks that entry-level people, if you all think about the first jobs that we had, or the first internships that we had, those are relatively confined, commoditized, you know, clearly described tasks, and we know that AI is really good at these commoditized, neatly defined tasks.
00:07:06 Fabian Stephany
But in the long run, it might ultimately
00:07:08 Fabian Stephany
backfire, right?
00:07:09 Fabian Stephany
Because the people, the junior people today are the mid-level management and the senior management people in 5 to 10 years.
00:07:15 Fabian Stephany
So I'll be cautious when it comes to saying like, okay, you can automate away these roles for various reasons.
00:07:21 Fabian Stephany
The other thing that we see in research where we are, I guess, at the forefront of this at the OII with our latest study is that we do see that the prospects of finding a job are significantly higher for people who have AI skills, where we
00:07:34 Fabian Stephany
to keep things short, we ran an experiment with recruiters around the question of like, who would you invite to an interview from the UK, Germany, and the US?
00:07:41 Fabian Stephany
And essentially what we see is that if people have AI skills on their CV, and particularly if they have a certificate on AI skills, I hope we can talk about this maybe later in this episode, we see that the chances of being invited to the interview are significantly higher.
00:07:54 Fabian Stephany
And the most interesting thing for me as an author of this study is that this is also apparently compensating disadvantages of age or lower levels of formal education where people are sometimes even
00:08:04 Fabian Stephany
have higher chances than their higher educated counterpart or their younger counterpart in this kind of experiment.
00:08:10 Prateek Sibal
And actually, at UNESCO, we are looking at this from the public sector side of things.
00:08:15 Prateek Sibal
And we are trying to understand how are jobs changing in the public sector, what are the new skills you need in the public sector to 1st govern AI and manage digital transformation more generally.
00:08:26 Prateek Sibal
And here we are seeing the emergence of a new kind of a profile.
00:08:30 Prateek Sibal
So you don't necessarily need to be a technical expert on AI.
00:08:35 Prateek Sibal
But what you need probably is a combination of skills, some bit of technical knowledge.
00:08:40 Prateek Sibal
So if you're procuring an AI system, you at least know to ask the vendor what is the accuracy of the AI system that you want.
00:08:48 Prateek Sibal
Then you also need to think about service design and delivery.
00:08:51 Prateek Sibal
So how do you integrate AI in the whole public-centered service that you're developing or even an internal service?
00:08:58 Prateek Sibal
A lot of times, this aspect of the users is totally missed out.
00:09:02 Prateek Sibal
And this leads to basically public services being developed and never used, or it's a real pain to use them.
00:09:08 Prateek Sibal
And the third aspect that we are seeing in terms of skills is around AI governance.
00:09:12 Prateek Sibal
So are you able to do an ethical impact assessment of an AI tool or not?
00:09:17 Prateek Sibal
are numerous examples where governments have used AI tools.
00:09:21 Prateek Sibal
I think there was a municipality in Norway somewhere which decided to close schools because they used a gen AI tool which hallucinated and provided fictitious evidence.
00:09:32 Prateek Sibal
So there are serious repercussions of using AI in the public sector.
00:09:36 Prateek Sibal
So you need to prepare your managers with the right kind of governance skill for this technology.
00:09:42 Prateek Sibal
So if you have people who have
00:09:44 Prateek Sibal
these three aspects, the technical, service design, governance, skills, I think there would be better managers in the public sector, and your quality of services would also improve.
00:09:55 Prateek Sibal
So this is a trend that we are seeing in some of the research that we are doing at UNESCO.
00:09:59 Veena McCoole
Tell me more about this topic of micro-credentials.
00:10:01 Veena McCoole
It seems to be an interesting and relatively accessible short-term way that people can get upskilled in an industry that's moving very quickly and that they may not necessarily have the resources or bandwidth to necessarily pursue a larger, more traditional academic degree.
00:10:18 Prateek Sibal
I mean, for us, again, the entry point was for the public sector because we realized that
00:10:24 Prateek Sibal
working on AI governance since the past few years.
00:10:27 Prateek Sibal
Governments are being asked to do more and more, do better procurement, manage and govern AI systems, develop AI tools.
00:10:34 Prateek Sibal
But at the end of the day, it's civil servants who are supposed to do this job, but they're not supported to actually acquire those skills, but are being asked to do more and more.
00:10:43 Prateek Sibal
So we worked actually with the Said Business School here at the University of Oxford to develop a massive open online course on AI and digital transformation for government.
00:10:51 Prateek Sibal
And the idea here was to take some of these competencies that I talked about and turn them into practical tasks that people can do using AI.
00:11:00 Prateek Sibal
So if you are a communications person in government, how can you use AI to simplify a notice that you're going to publish a newspaper?
00:11:08 Prateek Sibal
Instead of saying, you are hereby informed and so on, but how to make it more kind of interactive and easier for people to engage with government?
00:11:17 Prateek Sibal
There are numerous tasks like this that we try to simplify across the
00:11:21 Prateek Sibal
The work cycle of a government official through this course, and there, I mean, it's the uptake has been really encouraging.
00:11:29 Prateek Sibal
We have, as we speak since about past four months, we have 30,000 civil servants from 192 countries who are taking this course, and we have demand for more governments saying...
00:11:41 Prateek Sibal
I want to make this course available in Hungarian or in Khmer or in Ukrainian, because they really see the value of this 12-hour-long learning program and how it really is designed for this specific audience and it changes the way they do their work.
00:11:57 Fabian Stephany
If I hear you say 12 hours, you might think, if you're listening to this, you might think, okay, 12 hours, how are 12 hours ever going to change my job prospects or the things that I learn about artificial intelligence or how to improve my job with AI?
00:12:08 Fabian Stephany
I think actually this makes a tremendous difference.
00:12:11 Fabian Stephany
I mean, looking at the study that I mentioned earlier, we see the real differences in terms of whether AI skills boost your chances of being invited to an interview.
00:12:20 Fabian Stephany
They often kick in when people have certain certificates.
00:12:23 Fabian Stephany
And from an economist point of view, it's pretty clear.
00:12:25 Fabian Stephany
It's kind of signaling theory, particularly for the cases where we had these disadvantages designed.
00:12:31 Fabian Stephany
So we had candidates that had a master's degree competing against candidates who did not have a master's degree.
00:12:37 Fabian Stephany
And of course,
00:12:38 Fabian Stephany
If we didn't change anything about the scenario, we see that the people with a master's degree have, on average, always the higher chances of being invited because a master's degree sends a signal.
00:12:46 Fabian Stephany
people learn something at the university, they have a credential for what the skills that they acquired.
00:12:52 Fabian Stephany
Once we add AI skills for the people who are disadvantaged, in a sense, who do not have the master's degree, it lifts the, it changes the odds in their favor slightly.
00:13:02 Fabian Stephany
But once we add a certificate, and that could be one of these courses, you know, 12, you said 12 hours, right?
00:13:07 Fabian Stephany
12 hours.
00:13:08 Fabian Stephany
maybe also slightly longer, a week or two weeks, they're not only at the same level, they are in some cases even at a higher level of chances of being invited to the interview.
00:13:16 Fabian Stephany
And that's, it's pretty clear, they lack the signal, the formal signal previously, they now have a formal signal for AI skills, and then the tide shifts.
00:13:25 Prateek Sibal
And I think just beyond the AI aspect, it also signals curiosity, openness to learn and challenge yourself, which a lot of people in the workplace are looking for because technology changes is quite rapid and people want
00:13:38 Prateek Sibal
a workforce that is committed to learning and growth.
00:13:41 Prateek Sibal
And not everyone can go back and do a master's every few years.
00:13:44 Prateek Sibal
So micro-credentials are a great way for people to upskill and signal that to the market.
00:13:49 Fabian Stephany
And also the question is like, is a master's degree the right kind of vehicle to do that, right?
00:13:54 Fabian Stephany
Because if you think about a master's degree, one-year master's degree, if I think about our students, of course, they get great technical knowledge in our master's program at the OII, but they often, you know, it's very time-intensive because they are building networks, they're meeting people, you know,
00:14:08 Fabian Stephany
It's a life experience.
00:14:10 Fabian Stephany
While for some of the things that you described, it's more relevant probably to have these technical skills, right?
00:14:15 Fabian Stephany
To know what AI can do and know about the limitations of this technology.
00:14:19 Fabian Stephany
And I think this is something that can very efficiently be delivered in the course of, for example, a day long or a week long or maybe month long class.
00:14:28 Fabian Stephany
Yeah.
00:14:29 Veena McCoole
So far this conversation is making me think about where the onus lies here.
00:14:33 Veena McCoole
Is it on me as an employee or
00:14:37 Veena McCoole
a professional to kind of, as you say, signal to the market by enrolling myself in a course like this?
00:14:42 Veena McCoole
Or is it actually someone else who's responsible for closing the skills gap at work?
00:14:47 Veena McCoole
I'm curious what you think about that.
00:14:49 Veena McCoole
You know, is it down to the individual?
00:14:51 Veena McCoole
Is it down to employers, governments in your case, Pratik?
00:14:54 Prateek Sibal
So I think individuals have a role to play, to upskill, but at the same time, institutions need to create that enabling space.
00:15:01 Prateek Sibal
When you're talking about technology adoption, let me take the example of the judiciary.
00:15:06 Prateek Sibal
We work with them quite a bit and we did a survey where we asked them, how are you using AI?
00:15:11 Prateek Sibal
This was a, we have a network of 36,000 judicial actors around the world.
00:15:15 Prateek Sibal
And they came back, about 44% of them said that they are using AI on a daily basis, some for search, some for legal drafting and so on.
00:15:25 Prateek Sibal
But only 9% had received any kind of institutional guidance or training to actually do so.
00:15:30 Prateek Sibal
And then there are a lot of cases where you have, for instance, in Colombia, a judge used ChatGPT to decide whether a child should get disability benefits or not.
00:15:41 Prateek Sibal
What this does is this lack of institutional guidance is that you have a certain set of people who are under extreme work pressure and they want to use AI and what we are calling shadow AI use, even if it's not authorized to get the job done.
00:15:54 Prateek Sibal
And then there are at the other end people who are totally scared because there is no guidance and they don't want to touch an AI system.
00:16:01 Prateek Sibal
In both cases, the institution loses.
00:16:03 Prateek Sibal
In the former, you lose because
00:16:05 Prateek Sibal
Well, you could actually make a lot of mistakes and trust in, for instance, the judiciary would decline if a judge is doing that.
00:16:12 Prateek Sibal
In the latter, there are some opportunities of using AI to make your life easier as a judge.
00:16:18 Prateek Sibal
You are probably not using it.
00:16:19 Prateek Sibal
So here, the institutions need to step in and provide guidance on how to use AI.
00:16:24 Prateek Sibal
But even beyond that, I think institutions have a key role to, staying with my example,
00:16:30 Prateek Sibal
of the judiciary, a judge is not going to sit and do an algorithmic impact assessment or an ethical impact assessment on an AI tool.
00:16:38 Prateek Sibal
You as an institution need to have the resources and the department or whatever people to actually do that.
00:16:44 Prateek Sibal
And then the judge is supported and say, okay, I can use this tool.
00:16:48 Prateek Sibal
and this is authorized by my institution and so on.
00:16:52 Prateek Sibal
So I think it's a two-way space where the individuals have a role, but institutions have a mega role to play, which is often not very clearly thought about in institutions.
00:17:03 Fabian Stephany
Yeah, and I fully agree with this.
00:17:05 Fabian Stephany
If you think about it, maybe that's a good point to...
00:17:07 Fabian Stephany
to advertise the book a little bit that I wrote.
00:17:10 Fabian Stephany
But frankly speaking, if you zoom out a little bit, we talk about AI here and the mismatch, the labor market mismatch in terms of the skills that are being required or increasingly required on the labor market, the skills that people have is already big when we only speak about this aspect.
00:17:24 Fabian Stephany
But this is just one of multiple changes that we're seeing on the labor market.
00:17:27 Fabian Stephany
To add a few more, it would be greening of the economy.
00:17:30 Fabian Stephany
You know, we need more solar panels, we need more refurbished homes.
00:17:33 Fabian Stephany
retiring baby boomers.
00:17:35 Fabian Stephany
There are 10s of thousands of baby boomers retiring every week in the UK or in the United States, for example.
00:17:40 Fabian Stephany
And this changes the labor market composition tremendously and widens the skill game even more.
00:17:45 Fabian Stephany
And if you look at these numbers, you immediately realize it cannot be only on the individual or only on the institutions.
00:17:52 Fabian Stephany
Everybody has to play a role if we want to manage this probably biggest challenge, at least from my point of view, biggest challenge, economic challenge that we have right now in front of us.
00:18:02 Fabian Stephany
from the individual over to schools, to universities, to governmental institutions.
00:18:06 Fabian Stephany
I think the case that you made here with the judiciary is a really strong one.
00:18:10 Fabian Stephany
Yeah.
00:18:11 Prateek Sibal
And I think beyond just talking about specific technology, we often forget in digital transformation that there's a transformation aspect of things, which requires a cultural change in institutions.
00:18:23 Prateek Sibal
You need a strong level of leadership support.
00:18:25 Prateek Sibal
And we did a study where we saw that
00:18:28 Prateek Sibal
about 51% of chief information officers in government said, well, we are actually blocked because the decision-making processes are so siloed.
00:18:37 Prateek Sibal
And one department is doing something and the other is doing something else, and there's absolutely no coordination.
00:18:43 Prateek Sibal
These are really institutional problems, which technology is not going to come and fix.
00:18:48 Prateek Sibal
Maybe you can be helped with better communication tools, but you need to think deeply about the design of these institutions.
00:18:56 Prateek Sibal
and how they organize workflow and processes.
00:18:59 Prateek Sibal
Without that, going back to the first point, you will have a lot of AI use, but it will probably not lead to productivity improvement because it is for low-hanging things where you're maybe doing a spell check or correcting the text.
00:19:14 Prateek Sibal
but not really for a wholesale reorganization of tasks to free up more time also.
00:19:19 Fabian Stephany
Or it might lead to improvement for some people, but there might be a majority even of people that are left behind because it requires these institutional changes, but also kind of a change of mindset.
00:19:28 Fabian Stephany
You know, is learning something that I aspire throughout my life, right?
00:19:33 Fabian Stephany
There's still many people in the labor market who've grown up in a time where you had this so-called front-loading in learning.
00:19:38 Fabian Stephany
You go to school, go to university, you finish with a degree eventually at, I don't know, the age of, in your mid-20s.
00:19:44 Fabian Stephany
And then you stick with the same job or maybe even the same company for the rest of your career.
00:19:48 Fabian Stephany
And there's relatively little learning going on.
00:19:50 Fabian Stephany
And I think if you look at this with a sober perspective, it's clear to all of us that those times are over, right?
00:19:56 Prateek Sibal
Absolutely.
00:19:57 Prateek Sibal
And here we see that the incentive structure is a key space to think about in the public sector, because it's exactly this mindset that, okay, I am now a civil servant.
00:20:07 Prateek Sibal
I have this job for life.
00:20:09 Prateek Sibal
What do I do?
00:20:10 Prateek Sibal
you are not really motivated to invest in upskilling.
00:20:14 Prateek Sibal
You have a lot of supply side interventions, you have courses, trainings available, but why do you do it?
00:20:20 Prateek Sibal
So I think governments are now also thinking about how to incentivize people, whether some places it's about creating digital units and then picking up within government people who have a certain set of digital skills, AI literacy, to kind of join those units.
00:20:37 Prateek Sibal
It is, again, a signal.
00:20:39 Prateek Sibal
And for that, you need to think of the incentive structure in government and the competency frameworks that they have in government, which need to be updated to reflect what changes AI or digital transformation afford in our civil services.
00:20:54 Veena McCoole
Okay, so what I gather there is that it takes both sides.
00:20:57 Veena McCoole
It takes...
00:20:58 Veena McCoole
yes, the willingness and the appetite to reshape one's own skills landscape as an employee, maybe some incentives in place, but it also takes an institution setting the tone from the top down and in some cases, leading the way.
00:21:12 Veena McCoole
And I'm curious if we kind of move to practical advice here, like how do employers or institutions set the tone for this type of AI transformation and lead the way for their people?
00:21:22 Prateek Sibal
I mean, there are numerous examples here.
00:21:24 Prateek Sibal
First is really
00:21:25 Prateek Sibal
to understand from the people what are the tasks and the processes that they are involved in.
00:21:30 Prateek Sibal
Oftentimes, we've seen that there is a brilliant idea in the IT department, or some vendors are pushing a particular tool and say, okay, this is an imagined need that our people have, and then that tool is pushed in the organization.
00:21:47 Prateek Sibal
But the people who are actually working, they realize that, okay, they first have to now migrate to a totally different tool that probably they did not need, or it does not fully meet their expectations.
00:21:58 Prateek Sibal
So first part of any kind of digital transformation journey in, for instance, the public sector, anywhere else, actually.
00:22:05 Prateek Sibal
is to go to and talk to the people that you're working with and ask them, what are your needs?
00:22:09 Prateek Sibal
What are your pain points?
00:22:11 Prateek Sibal
I would even rely on some design thinking approaches and they'll say, what are the pain points when you are interacting with X, Y, Z service or product?
00:22:19 Prateek Sibal
How do you feel?
00:22:21 Prateek Sibal
Because that's the phase where you're doing the discovery and trying to find ideas and problems.
00:22:26 Prateek Sibal
This is not even like finding the solution.
00:22:29 Prateek Sibal
A lot of times we're just so fixated on, I have the solution,
00:22:33 Prateek Sibal
But maybe that's not the problem that you wanted to solve in the first place.
00:22:37 Prateek Sibal
So really first go and do that.
00:22:40 Prateek Sibal
Once you do that, then you start after this discovery phase, slowly defining what is the problem.
00:22:47 Prateek Sibal
And then you potentially build solutions.
00:22:49 Prateek Sibal
And this would in some cases look like maybe I've understood the problems and I need an AI strategy for my organization, link it with the different aspects of HR, upskilling and so on for the people.
00:23:02 Prateek Sibal
And then maybe you have some semblance of a digital transformation plan in place.
00:23:08 Fabian Stephany
Yeah.
00:23:08 Fabian Stephany
I think if I think about this question, I probably think about two things.
00:23:11 Fabian Stephany
One is skills and one is technology.
00:23:13 Fabian Stephany
And you could look at this from having a pessimistic point of view and say like there's increasing inequality in the world because there's some actors who very well understood, already understood, with the example of AI, to navigate around these two poles of skills and technology.
00:23:28 Fabian Stephany
many big tech companies do because they've built these tools in-house and they know both how to integrate the technology into efficiently along all of these the things that you said but dig into the workflows and they also understood how to keep their workforce afloat when it comes to skills.
00:23:44 Fabian Stephany
I don't want to take this pessimistic point of view.
00:23:46 Fabian Stephany
I'd rather like to see this as an opportunity because this shows us there are best practices that work, right?
00:23:51 Fabian Stephany
Both on a company level, also on a government level.
00:23:54 Fabian Stephany
There are governments, usually, you know, the typical digital leaders like Estonia, for example, that manage to move ahead with regard to both technology and skills.
00:24:04 Fabian Stephany
And I think often it comes back when things do not work, it comes back to a lack of guidance, for example.
00:24:09 Fabian Stephany
Companies, particularly small and medium-sized enterprises,
00:24:12 Fabian Stephany
overestimate the costs of having a system in-house, like an AI system in-house, because of course there are privacy concerns, there are data breach concerns, there are also environmental concerns, and that is often due to these concerns are kind of exaggerated because people, for example, overestimate the costs.
00:24:28 Fabian Stephany
They think it costs much more to install the system in-house than it actually does.
00:24:32 Fabian Stephany
And the other one is guidance, right?
00:24:34 Fabian Stephany
Often when we talk to people in companies who mentioned the shadow use of AI, so employees not using artificial intelligence or secretly
00:24:41 Fabian Stephany
using artificial intelligence because they don't have the guidance.
00:24:44 Fabian Stephany
Management doesn't say, we want you to use this.
00:24:46 Fabian Stephany
We actually not only want you to use this, we have a tool in-house that you should be using for X, Y, Z according to the skills and the problems that you have.
00:24:54 Fabian Stephany
And I think that's particularly for firms, the backbone of the economy in countries like the UK, but also the US and in Germany, the small and mediums as enterprises, to have this kind of guidance and lead the way, management leading the way for the workforce.
00:25:09 Veena McCoole
So if we kind of take this back to where we started today's conversation, we've covered so much here in terms of the public sector, in terms of the skills gap.
00:25:16 Veena McCoole
What does all of this mean for the future of skill building in the modern workforce?
00:25:23 Fabian Stephany
From my point of view, I mean, running a group here at the OI called Skill Scale, you're not surprised to hear me say that skills is probably maybe one of the most important ways to leverage all of this and help to shape a sustainable future.
00:25:36 Fabian Stephany
When it comes to skills, though, interestingly, we very, very often lack the first step, which is a map of where we actually are, right?
00:25:46 Fabian Stephany
What are the skills that I have?
00:25:48 Fabian Stephany
What are the dormant skills that I might be having, the skills from a previous job that I'm not using?
00:25:52 Fabian Stephany
And the reason why I'm saying this is because these skills crucially define your adaptability, right?
00:25:58 Fabian Stephany
Being able to maybe move to a slightly different role because of technological advancement, because AI is getting better in writing emails, summarizing reports,
00:26:06 Fabian Stephany
And if that had been a major part of my work, I obviously need to shift, and some of the skills that I had from a previous role might help in this.
00:26:14 Fabian Stephany
And so I think having a clear idea of the skills that people actually possess is kind of the, maybe not the sufficient, but definitely the necessary condition for moving ahead.
00:26:25 Prateek Sibal
And as I was mentioning about this competency framework, Fabian,
00:26:29 Prateek Sibal
And here what we tried to do was map out what are the skills needed in the public sector.
00:26:34 Prateek Sibal
But these were not only like skills, but also attitudes.
00:26:38 Prateek Sibal
So experimentation, openness to collaboration.
00:26:42 Prateek Sibal
These things are extremely important when you are working in teams and you want to get things done.
00:26:48 Prateek Sibal
And we are actually missing these attitudes in a lot of cases in the public sector.
00:26:54 Prateek Sibal
So we argue for first
00:26:57 Prateek Sibal
skills around basic data literacy, data-driven decision-making, AI literacy, but also around user-centricity.
00:27:05 Prateek Sibal
How do you develop an attitude that is user-first?
00:27:09 Prateek Sibal
How do you think in a more agile manner and not be stuck that, oh, I had a plan, and then I executed one, two, three, four, but you do it in an iterative way.
00:27:19 Prateek Sibal
You say, okay, I tried, I test with the users, and then I go back to the drawing board.
00:27:24 Prateek Sibal
And I think this
00:27:26 Prateek Sibal
This attitude, this kind of putting it in practice requires an ability to experiment.
00:27:33 Prateek Sibal
But at the same time, from the institution side, it requires acceptance of failure, which is not something that public institutions are designed for.
00:27:44 Prateek Sibal
If you fail, then basically you are held accountable for the damage because it's public money.
00:27:50 Prateek Sibal
But I think we are losing something, some innovation in the public sector.
00:27:54 Prateek Sibal
Absolutely.
00:27:54 Prateek Sibal
To even adapt, as you were saying, an AI tool and develop it in-house requires a fair bit of courage to say, I will do it, and maybe there will be some problems, but let me try.
00:28:08 Prateek Sibal
And maybe in the medium to long term, it pays off to have an in-house tool in terms of privacy, data protection, and so on.
00:28:15 Prateek Sibal
So we need this experimental mindset also.
00:28:19 Fabian Stephany
Absolutely.
00:28:20 Fabian Stephany
This kind of openness is probably one of the most, on the individual end and on the organization, most crucial things, generally speaking, but particularly in a time where we are right now, because things are so uncertain, right?
00:28:33 Fabian Stephany
If you talk to the people of the, again, the institutions that are often perceived being very successful, big tech companies, for example, mastering this change,
00:28:42 Fabian Stephany
That is one of their crucial paradigms.
00:28:44 Fabian Stephany
It's like there is no failure in that sense, right?
00:28:47 Fabian Stephany
It's like whenever, or let me put it differently, whenever you try something, you will definitely fail.
00:28:52 Fabian Stephany
That's part of the process.
00:28:53 Fabian Stephany
And that is the huge opportunity to learn something.
00:28:55 Fabian Stephany
That is, it is an indicator of making progress.
00:28:58 Prateek Sibal
As Nelson Mandela said, it's either you succeed or you learn.
00:29:04 Fabian Stephany
Yeah.
00:29:04 Veena McCoole
And what I love about what you've both shared there, you know, Fabian, you talked about
00:29:09 Veena McCoole
skills that can adapt and transfer even between jobs.
00:29:12 Veena McCoole
Pratik, you talked about having a willingness to experiment and be open.
00:29:17 Veena McCoole
These are inherently human skills that only we can bring to our jobs.
00:29:22 Veena McCoole
Yes, we're talking about AI, but we're also talking about the almost sharpened role and the responsibilities that we need to take as human employees in today's day and age.
00:29:31 Fabian Stephany
Yeah, absolutely.
00:29:32 Fabian Stephany
I mean, ultimately, I mean, we're talking about AI today.
00:29:35 Fabian Stephany
in five years from now, we might no longer be talking about AI because it's like become like electricity or the computer to us in the sense like a general purpose technology.
00:29:44 Fabian Stephany
But I would, I don't want to make too many forecasts because I know forecasts in economics notoriously age badly, but we will be talking about change in the future again.
00:29:53 Fabian Stephany
And this is essentially a conversation that we're having about change.
00:29:56 Prateek Sibal
Absolutely.
00:29:57 Prateek Sibal
I mean, if you, we are all knowledge workers, right?
00:30:00 Prateek Sibal
And if you are not curious and not trying different things and experimenting,
00:30:05 Prateek Sibal
then what is the point?
00:30:07 Prateek Sibal
You don't innovate.
00:30:08 Prateek Sibal
So those attitudes that you mentioned, Vina, are extremely important for us to advance in our careers.
00:30:15 Veena McCoole
And my last question for both of you as we wrap this up is, you know, what's one thing you'd like the listener who is facing this AI jobs market to take away in terms of skills, in terms of how they run their companies?
00:30:28 Prateek Sibal
All of us feel a lot of pressure and anxiety about artificial intelligence and it taking our jobs.
00:30:34 Prateek Sibal
And
00:30:35 Prateek Sibal
We've seen whatever little evidence that we have that it has not reached the level where it is so integrated that it is leading to mass layoffs and so on.
00:30:44 Prateek Sibal
But there is a whole narrative around it that is extremely anxiety-inducing.
00:30:50 Prateek Sibal
So first, we need to calm down and reflect.
00:30:55 Prateek Sibal
What we need to do next is really, I think, look at what aligns with your own skills that you have currently.
00:31:02 Prateek Sibal
then try to see what complements those skills and try to branch out and develop those skills.
00:31:06 Prateek Sibal
As you were saying, Fabien, what are the complementarities that I can bring?
00:31:10 Prateek Sibal
As an individual, I always think about if I have expertise in one domain and I want to venture out into a new domain, I think about what do I bring to this adjacent domain while learning about
00:31:24 Prateek Sibal
the structure and the knowledge of that particular domain.
00:31:27 Prateek Sibal
So I think that's a good place to start while you're finding whatever AI courses you're looking at, not necessarily go deep into, oh, I want to tomorrow start coding or try to build AI apps.
00:31:40 Prateek Sibal
Just look at where your skills are and then link it with the adjacent field where you want to grow in.
00:31:46 Fabian Stephany
I think my take would be to say AI is just one of many examples where we see
00:31:53 Fabian Stephany
change unfolding rather rapidly.
00:31:55 Fabian Stephany
And that's also what I'm saying in Skilled in the book is that where I try to give many examples of change, very challenging situations like the pandemic, for example, or challenging situations like, you know, we need to quickly transition into more sustainable, ecologically more sustainable economy.
00:32:13 Fabian Stephany
And this is rapid change.
00:32:15 Fabian Stephany
This can be very stressful, particularly if you are guided by, from my point of view, false narratives of dystopian futures.
00:32:22 Fabian Stephany
but it's also full of opportunities.
00:32:24 Fabian Stephany
There are countless examples for people with the right mindset and the right skills that excelled not despite of this tremendous change, but because of this tremendous change.
00:32:34 Fabian Stephany
And that's also certainly true for the age of AI.
00:32:37 Veena McCoole
Thank you so much, Fabian and Pratik, for joining us on this episode of The Human Interface.
00:32:42 Veena McCoole
Make sure to subscribe to be alerted when new episodes drop on Spotify, Apple, or anywhere else you listen to podcasts.
00:32:49 Veena McCoole
and follow the Oxford Internet Institute on LinkedIn and Instagram to stay abreast of our latest research and updates.