This is an audio transcript of the Working It podcast episode: ‘Bored at work? How AI could come to the rescue’

Jeff Wong
So what we see is more productivity, being able to answer more questions for clients and customers and frankly, what the teams love the most is that they’re able to spend their time thinking more using the human part of their brain. For us, it is a booster for us in every way possible.

[MUSIC PLAYING]

Isabel Berwick
Hello and welcome to Working It from the Financial Times with me, Isabel Berwick. The hype surrounding generative AI and ChatGPT in the workplace is now fast becoming a reality. Some recent studies suggest that up to two-thirds of occupations are exposed to AI to some degree, and a bot could do between a quarter and a half of the workload. So the question today is, will AI make some jobs redundant or will it simply change the way we work? Jeff Wong, who you just heard at the top of the show, heads up the innovation team at the professional services firm EY. They’re looking at how AI could transform their work and we’ll hear more from him later. Plus, how worried should workers be? And are regulators starting to get involved? But first, to remind us what this technology is all about I spoke to Madhumita Murgia. She’s the FT’s artificial intelligence editor. Madhu, welcome to Working It. We’re talking about AI so much now in connection with professional and knowledge-based jobs. And I keep hearing these terms like “large language models” and “generative AI”. Could you explain to listeners what do these terms actually mean and where are we now with AI in the workplace? It seems to change every week.

Madhumita Murgia
(Laughter) Yeah, I know, and it’s quite hard to keep up with, even though I’m looking at it all the time. So I think AI is becoming this umbrella term. It’s artificial intelligence. What it really means is software that needs tons and tons of data to be trained on and then it’s able to draw patterns out of that data to come up with some kind of output. So the output could be words, it could be numbers, it could be images. We already had AI systems embedded into finance and banking. We had it embedded in, you know, HR and recruitment tasks because essentially they’re just technologies that can find patterns that humans may miss when they look at large amounts of data. So it’s already been in the workplace in that way. Now, generative AI is something, is a term from kind of the last year or two, and what it refers to is software that can make things, that can create things. So that could be either images or it could be text. So, many people may have heard of ChatGPT, which is essentially what you called a large language model. And what that really means is it’s a software that has been trained on billions of words or fragments of words taken from the internet, and it means that it can now generate words and sentences and paragraphs and essays in a way that sounds really plausible and, you know, really nuanced, even.

Isabel Berwick
Where do you see the biggest shifts actually beginning to happen right now in the workplace?

Madhumita Murgia
So, I would say the kind of the big moment, which will affect many millions of people right now, is AI or generative AI being integrated into the software that we use daily in the workplace. So things like Microsoft Office, Google Docs, Gmail, Outlook — these kinds of products that we all use. Both Microsoft and Google, who are obviously the giants in the productivity software space, have announced plans to integrate AI into these products. And what that will mean is AI predicting what your email should look like for example in Gmail, or suggesting what a document could look like or generating a document for you when you just give it a little prompt. PowerPoint, Excel as well. These big companies have been key figures in developing AI software as well; Microsoft more from an investment perspective because they are the people who invested, the biggest investors in OpenAI, which makes ChatGPT. And Google, of course, has been a pioneer in the AI space for many years. So I think that will be really interesting. It’s gonna touch all our lives in quite an immediate way.

Some of the things that I find really interesting in terms of jobs and other more creative jobs, traditionally you think of technology automating sort of more manual jobs because they’re stronger or more efficient than humans. But really what we’re seeing with generative AI is that it’s impacting all of these professions that we always thought were the purview of humans. There was a piece recently that I saw about a whole group of video game designers in China who admitted, yes, it made them more productive, but also so much more exhausted, and also that there was the sense of the creativity being taken away from them. There’s also areas like voice actors, where you can use AI to mimic voice. And then, you know, of course, since ChatGPT writes text and can translate as well, you’re looking at interpreters, translators and those of us who write for a living as well. So I think it’ll be really interesting to see that kind of impact on the creative areas.

[MUSIC PLAYING]

Isabel Berwick
We’ll come back to Madhu again later. So let’s hear from a company already making use of the technology. I spoke to Jeff Wong. He’s global chief innovation officer at the professional services firm EY. I asked him for some examples of how his company is currently making use of AI in its processes.

Jeff Wong
We actually are pretty excited about something we’ve announced publicly. It’s called our EY Intelligent Payroll Chatbot, and it’s where we’re actually going through Microsoft to the GPT platform, OpenAI’s GPT platform, and we created a chatbot around our payroll services. And what’s really interesting is the statistics around how much, at least in the pilot phase, how much better we are at answering questions early and quickly and accurately has gone up. But also then the speed at which we’re able to address these questions, these complex sort of employee-payroll questions, it’s been pretty remarkable. So payroll taxes is a great example, our chatbot where people can ask a multitude of different questions. How does this tax work? In my country, they can do it in a multitude of languages. Currently, our pilot chatbot does it in 27 languages that we’re comfortable saying it’s at least 27 that it can do it well. We have hundreds of these types of proposals. What we’re trying to do is be very thoughtful about going after the best opportunities, but also being very thoughtful about the responsibilities around that. Does it give accurate information? Is it honest and truthful for what we’re trying to have it do and we’re measuring? Does it really have the impact that we want? So if there’s a significant impact that we’re looking forward to from this application, this technology, but we’re really being thoughtful about our responsibility here as well.

Isabel Berwick
How do you envisage it working in practice?

Jeff Wong
I think what’s really interesting within professional services is generative AI actually proposes — I’m not saying it will do this, but proposes — that we can rethink the entire set of processes themselves. So instead of going through the normal set of processes that we go through to ask, for example, if there’s anomalies in the data, right where we go through a normal set of steps: you know, download the data, move it here, move it there, transform it, chart it this way, etc. It looks like it’ll help improve us going through those steps faster, but it also allows the possibility that we could simply have the data available to the generative AI effort and simply ask the system directly. Is there an anomalous factor in this data? And that would just jump over some other process that we had before. And I think that that’s really what makes generative AI disruptively fascinating and interesting, particularly when it comes to, I guess, what people would think of traditional white-collar work, right?

Isabel Berwick
So talking of say, you know, if half if your job has gone or has been freed up, do you still have a job or what are you doing with extra time? It seems like an existential question that we haven’t really thought about yet.

Jeff Wong
Isabel, that’s a very good question; I think the right question. So what we see is more productivity being able to answer more questions for clients and customers. And frankly, what the teams love the most is that they’re able to spend their time thinking more using the human part of their brain. For us, it is a booster for us in every way possible.

[MUSIC PLAYING]

Isabel Berwick
So Jeff Wong at EY sees plenty of positives, but his enthusiasm isn’t shared by everyone. A growing number of AI experts have been ringing alarm bells about the fast rollout of powerful large language models like ChatGPT. So I asked Madhu how she sees the increasing use of AI having an impact on the jobs that many of us are doing now.

Madhumita Murgia
I think there’s, you know, the doomsday AI and then the sort of very happy clappy. This is gonna be amazing for GDP and productivity. And then there’s a whole range of people in between. I fall in the category where, of course, you think about which parts of your job could be automated and then think about what would that free you up to do. So I think the big thing going forward now in workplaces will be how leaders manage this within their organisations. Are they gonna talk to people about what they’re experimenting with and trying out and kind of reassure people that, you know, they will be using this in some ways to help them and to improve their productivity, but in other ways will leave their sort of unique creative contributions untouched? So I think the big thing going forward will be how that’s communicated to people so that everyone doesn’t panic. And also really to allow people to experiment with making their jobs easier and more efficient. I think giving people permission to experiment with it on their own means that employees have more agency, right, in how they choose to play with it. And then we can figure out ourselves, you know, really, how can it help and where it isn’t very good. If you allow people to play with it, then we can also figure out the limits of it within our work and also see how we could help ourselves be more productive. And that kind of, I think, helps with the fear of it.

[MUSIC PLAYING]

Isabel Berwick
The fear of what AI could do to jobs is already very real across a number of industries. Mary Towers is a policy expert on employment rights. She works for the UK’s Trades Union Congress. It’s an umbrella body for unionised labour. Mary’s been researching how AI affects work for the past three years and she thinks the key issue is how much control workers will have over the way it’s introduced to their workplace.

Mary Towers
There’s change being imposed on people or as an act of collaboration, are the different groups in society being actively involved and consulted on that change? And if they are, then that change is going to be for the good and it’s going to represent all of the interests at work. I think another thing that I would highlight that is very different in this particular context is issues of scale. Because if let’s say you have a discriminatory algorithm that is making discriminatory decisions about people at work, for example, if you’ve got facial recognition technology that’s operating at work that only recognises white faces and not black faces, or that in some way gives lower ratings where you’ve got a disabled person with perhaps a facial disfigurement or different ways of expression. If you’ve got that type of discrimination algorithm in place, it’s going to have been rolled out at scale beyond just one workplace. You know, potentially, if you’ve got an international organisation, that algorithm might be operating across thousands of different offices.

Isabel Berwick
And how do you think AI can be helpful for workers? You know, what’s the upside here?

Mary Towers
There are all sorts of opportunities in terms of making work more rewarding. So removing laborious tasks, removing tasks that are potentially unsafe, for example. So there are huge opportunities there. But also we say there are huge risks in terms of potential discrimination, in terms of intensifying work to a level that is unsustainable and results in risks to both mental and physical wellbeing of workers.

Isabel Berwick
And so you’re looking at it from a UK perspective. But, you know, are you working with European regulators or with American regulators and unions? How is this shaping up globally? Because this is not a national problem, is it?

Mary Towers
There’s a huge risk that the UK is going to become an outlier in these issues. At the moment there are really significant steps that are being taken in other countries to ensure that AI is properly regulated. An example of this is the EU AI Act, a piece of legislation that provides protection and regulation in relation to the use of AI in all different aspects of society, not just work. Whereas in the UK, the government in its recent white paper suggested that there was no intention to regulate in this space and that instead there would be use of a set of ethical principles to be enforced by regulators who are already under-resourced.

[MUSIC PLAYING]

Isabel Berwick
Even in the time we’ve been recording this program, AI has changed. It’s in the news almost every day. It’s already here and some jobs will likely be lost or will change profoundly, and we just don’t know how exactly that will play out. But I guess the best-case scenario is it’ll take away the boring parts of our jobs and free us up for more human connection or more creative tasks. It could really unleash us from the process-driven boredom of work. And that’s incredible. But it could also lead to millions of people losing their jobs. Regulation’s needed urgently. And from what Mary Towers told us, it sounds like the US and the EU are leading the way. But if AI could take 50 per cent of the tasks, some of those jobs are going to go. And I also think there’s gonna be a coming massive resistance to AI. New research from Pew suggests American workers overwhelmingly reject the idea of using AI for making, for example, final hiring decisions by a margin of 10 to 1. So as AI enters workplaces in the mainstream, are we going to see a lot of worker resistance? I think so. But as Madhu says, it’s up to companies to prepare workers for this future and educate people to work alongside AI. And I guess that’s the next step in our journey.

[MUSIC PLAYING]

My thanks to the FT’s Madhumita Murgia, Jeff Wong at EY and Mary Towers at the TUC for this episode. If you’re enjoying the podcast, we’d really appreciate it if you left us a rating and review on Apple Podcasts. And please do get in touch with us. I’m at isabel.berwick@ft.com or find me on LinkedIn. If you’re an FT subscriber, please sign up for the Working It newsletter. We’ve got the best workplace and management stories, plus my office therapy advice column. Sign up at FT.com/newsletters. This episode of Working It was produced by Audrey Tinline. The executive producer is Manuela Saragosa with mix from Jake Fielding. Cheryl Brumley is the FT’s global head of audio. Thanks for listening.

Copyright The Financial Times Limited 2024. All rights reserved.
Reuse this content (opens in new window) CommentsJump to comments section

Comments

Comments have not been enabled for this article.