The hottest topic in the industry at the moment is, without doubt, Artificial Intelligence. At some point during the discussion in pretty much every podcast interview I’ve published in the last 12 months, we talked about the implications and use-cases of AI in recruiting. But, taking a step back, do we really understand what Artificial Intelligence is and do we have enough understanding in the industry to sensibly debate its implications and use-cases.
To help me answer these questions, I’m delighted that my guest this week is Rob McCargow, AI Programme Leader at PwC and Advisory Board Member for the All-Party Parliamentary Group on Artificial Intelligence.
In the interview we discuss:
• The definition of Artificial Intelligence and why it is such a hot topic
• Cutting through mythology and hype
• Applications, use-cases and risks
• The dangers of biased data
• Potential workforce displacement and what employers should be doing now to deal with it
Rob also shares his thoughts on AI’s likely effect on recruiting and talks about the adjustments PwC are making to their workforce strategy.
Subscribe to this show in Apple Podcasts
Transcript:
Matt Alder [00:00:00]:
Support for this podcast comes from the Innovation Circus, the latest experience from the Recruitment Events company. Hosted in partnership with Broadbean. The event takes place on 11 May at the British Library in London and this year the focus is on social recruiting and future innovations. Speakers from Google, Facebook and Vodafone amongst others will be on hand to share their stories and offer real life case studies for the over 200 in house recruiters in attendance. Plus two workshops will offer delegates a more hands on experience to solve their real life challenges. Go to www.innovationcircus.co.uk for more information and if you use the Code podcast you’ll receive a 20% discount on your ticket. That’s www.innovationcircus.co.
Matt Alder [00:01:18]:
Hi everyone, this is Matt Alder. Welcome to episode 125 of the Recruiting Future podcast. The hottest topic in our industry at the moment is without doubt artificial intelligence. Pretty much every podcast interview I’ve done in the last 12 months at some point during the discussion we’ve talked about the implications and use cases of AI in recruitment. But taking a step back, do we really understand what artificial intelligence actually is and do we have enough understanding in the industry to sensibly debate these implications and these use cases? To help me answer the question, I’m delighted that my guest this week is Rob McCargow, AI program leader at PwC. Rob is a real subject matter expert in this field and it’s brilliant to have him on the show.
Matt Alder [00:02:12]:
Hi Rob and welcome to the podcast.
Rob McCargow [00:02:14]:
Thanks Matt, thanks for having me along today.
Matt Alder [00:02:16]:
So could you introduce yourself and tell everyone what you do?
Rob McCargow [00:02:20]:
Yes. So I’m Rob McCargow, I’m the artificial intelligence program leader at PwC. I do a number of things so our team looks at both the application of AI to transform PwC from within, but also developing new services and products for our clients to solve their problems. I also get involved with a number of other activities. I’m on the All Party Parliamentary Group for AI’s advisory board and this is set up to help inform politicians of some of the policy implications of this technology. I also specialize in the way that the impact of the ethics of AI starting to unfold in business and society through a nonprofit called the IEEE Standards association and a number of the other areas involved with would be looking at the impact of our workforces with regards to automation led by technology such as AI and some of the Economic and commercial implications as well.
Matt Alder [00:03:25]:
Fantastic. So probably the best place to start is AI is a much talked about sort of topic in the HR and recruitment industry. In fact, you probably can’t go to a single conference without someone bringing it up. But sometimes I think there’s some sort of confusion about what it actually is. So could you sort of give us your definition of artificial intelligence?
Rob McCargow [00:03:50]:
Absolutely. So there is a lot of mythology and hype around what this technology actually is. The official definition is that it’s a branch of computer science dealing with the simulation of intelligent behavior in computers, or alternatively the capability of a machine to imitate and mimic intelligent human behavior. What this means in practice really is technology that’s been around for decades, 50, 60 years or so. That’s really starting to come to fruition now due to a number of key factors. First of all, we’re seeing an explosion in the curation and collation of big data. And this technology requires huge volumes of data in order to train the models. But also we’ve seen a proliferation in the power of computing power in the cloud. So that, coupled with both investment into the technology and the data and some advances in some of the technology, has seen a real amplification of its application to business problems.
Matt Alder [00:04:56]:
So you, you know, you do a lot of speaking, you do a lot of research, you’re, you know, you’ve got a view across a number of different industries. What are you seeing as sort of current applications of AI? How’s it working in practice at the moment? What might happen in the future?
Rob McCargow [00:05:12]:
We’re not seeing whole scale enterprise adoption of this yet. There’s still a lot of work being done in narrow proofs of concepts to apply to specific use cases. So across, for example, the HR function, there is technology starting to be applied to how do you augment the way that you select CVs? There’s some interesting breakthroughs in applying this to the way that video interviewing streamlined. Really the main growth in this technology has happened in the consumer space for quite some years. So whether this is in helping to augment the way that your next movie is found, or alternatively to help to shepherd you towards your meeting on time, really the focus now is how do you start scaling this in enterprise to achieve really meaningful business outcomes right on that tipping point where we’re starting to see some really promising breakthroughs in business. And this then brings both the opportunity to drive down costs, to drive up revenue, to manage risk better. But in itself it also does present a whole new category of different risks and considerations, businesses to contend with at the same time.
Matt Alder [00:06:31]:
That’s, that’s interesting. Could you give us a bit more insight into what some of those, some of those kind of risks might be?
Rob McCargow [00:06:37]:
There’s quite a few different categories, I think. First things first, if the technology scales as many my colleagues believe it will, it really does start to significantly change the essence of companies business models. It has a really transformative ability to disintermediate incumbents. It has the ability to really allow new entrants to scale very quickly to challenge the current market dominant players across all sectors. What the sort of risks we’re starting to see emerge, and there’s been some very well cited use cases in the consumer market, is when there’s an issue in particular around the way that the data to train the model is collated and curated. And what this means is that if you’re putting in biased data, the technology has the capability to scale and amplify that discrimination, which can amplify bias and actually lead to some quite harmful outcomes in business. One of the issues there in particular is the fact that because we all have unconscious bias in us, and the fact that the workforce training these systems is overwhelmingly male, there is a risk of gender discrimination coming through as one particular issue. Moving ahead though, one of the other issues that we will start to see is this technology first of all start to augment our capability, but then move to a capability to make autonomous decisions where there is not a requirement for a human in the loop to chaperone those decisions. And this of course opens up a whole new chapter of different business risks. One of the issues, first of all, is the displacement of existing workforce. One of the studies that we published recently suggests that by the early 2000 and 30s, up to 30% of the existing UK workforce could become highly susceptible to automation by technology such as AI. To put that in real terms, that’s 10 million workers in the current workforce. What we also see though, is because you get this substantial economic boost through the growth in productivity, and as you ever more personalize products and services, you see a consumption boost. It does give us a fairly sizable economic prize to aspire to. And with that, that then brings us the opportunity to think about adapting the workforce. Lifelong learning, reskilling and retraining to prepare people for a certain future that is hard to predict accurately. Beyond that though, there’s risks also with regards to the technology not being governed, risk managed appropriately. In particular, because the most advanced types of this technology are relatively opaque in how they make decisions, they can lead to remarkable breakthroughs. And optimal decisions. But the way that you can inspect the algorithm that’s making those decisions is sometimes quite a hard technical challenge. And when you start applying this in regulated industries, this can of course lead to distrust and potential damage. In particular, with the GDPR coming into force in the next few months, I.
Matt Alder [00:10:05]:
Mean, lots of, lots of interesting stuff there. And I think we could probably talk for hours and hours about, you know, privacy and bias data and, you know, all the, you know, all the significant sort of ramifications of this, but just to sort of, you know, zero in on the, on the, on the jobs part and the displaced workforce aspect of this. Obviously there are a couple of sort of broad views of the future. You know, one’s quite optimistic that saying new jobs will be created, and there’s obviously a pessimistic version that says, you know, lots of people will be unemployed, and there’s a hyper pessimistic version where we’ll end up living in some dystopia being ruled by robots. Where do you kind of sit in terms of the view of the future and also what do you think employers should be thinking about and doing now to prepare for it?
Rob McCargow [00:11:00]:
You’re absolutely right there. There is a real variety of perspectives and opinions on this. And I think that word perspectives and opinions actually sort of nails it, that there’s very little granular analysis of what might happen. And the fact is, is that the scare stories sell better and they sell better in the press. So, for example, you’ll often see a breakthrough in AI that has an impact on the workforce. And the word AI is replaced with robot. Cue the terminator. Arnold Schwarzenegger clip art to go into the piece in the press, which, as you can imagine, if you’re sitting there as a general member of society or in the workforce, it certainly does increase levels of anxiety. But there’s also the other side of the equation. The people are referred to as the tech utopians, the ones that would say, unleash the power of this technology, it will work itself out in the long run. And I think they’re also misguided in their approach. And we start seeing that happen with recent stories with regards to technology impact on society. So my view is, and I’m trying to sort of steer a calm, sober central course through this debate, is that we can be certain that there will be workforce disruption and displacement. It’s always happened. It’s happened across whatever you want to refer to them as the various industrial revolution stages. In the past, however, we have seen these technological breakthroughs eventually Leading to substantial growth, increase in wealth, opportunity and whole new categories of jobs that we never thought we’d have. Can we be certain this time? There is a risk that we look back on history as the sole guiding light to shape what the future might look like. Could it be different this time is a question that’s often asked because in the past technology has tended to displace manual labor. Could this in the future start displacing more cognitive labor, the very preserve of us as humans? And there are people that suggest that the. There’s concerns about technological unemployment, but beyond that, technological unemployability, which is a whole other concern for us. But I think you can only do with, as a business leader or an HR director, you can only contend with the facts and you’ve got to take a bit of a big leap into the future. One thing you do need to do though, I think is act now. This won’t necessarily unfold imminently, but the things you can start doing is starting to adapt your graduate intake. So for example, we know that our graduate intake will substantially change in the years ahead. So a few months ago we launched technology degree apprenticeships with three universities in the UK which are fully funded four year programmes for school leavers, where we give them a degree in computer science. And that sort of feeds our longer term workforce strategy. There’s also, I think, a number of things that business leaders and HR directors need to do as well. You can be certain back to what I said before, that a decent proportion of your workforce is equally excited and anxious about the impact of technology on their jobs. Now we can’t protect jobs, but we absolutely can protect people. So I think there’ll be an alignment in the years ahead where organizations become far more purposefully focused on their people. How do they equip them, how do they retrain and reskill them, but how do they communicate with them? The fact that change is coming but they’re going to attempt to be ahead of the curve and put their interests at the heart of it.
Matt Alder [00:14:59]:
No, there’s some, there’s some really interesting, really interesting stuff there and you know, obviously it’s something that should be on, you know, on, on the radar of every, of every business. Obviously you’ve got a view across a lot of industries, but I’m kind of interested to, to see what sort of perspective you might have on how AI might sort of transform recruitment or affect recruitment in the future. We talked about, you know, talked about unconscious bias and all that kind of stuff. But from your perspective, how do you Think AI is going to affect that particular industry?
Rob McCargow [00:15:36]:
I’m very optimistic about the application of this technology to AI. I mean, I might sound a bit downcast about it, but maybe I’ll just reframe it. I’m a responsible optimist about the application of this to HR recruitment. So what do I mean by that? I think the whole series of activities in the recruitment cycle that technology like this can make increasingly more fair, more equitable and more consistent in its ability to make decisions. I think about ourselves, you know, one of the largest recruiters in the country. If you’ve got teams of people spending all day on the applicant tracking system pressing accept or reject, there’s no doubt that cognitive dissonance and unconscious bias is evidently at work there. So this technology absolutely offers the opportunity to improve those sorts of processes. I always think of it on the other side of the equation though. If you just leave it to work its supposed magic, then you are going to be looking at some substantial risks of actually making the situation worse. So for me, I think we need to move on with this. We need to improve some of these parts of the recruitment and HR cycle, but we need to do this in a way that actually allows participation. It’s not the sort of thing that you just leave as a standalone IT project. And I think this is a great opportunity for HR and recruitment professionals to step up to become sufficiently adept in the technology, understanding the opportunities and the risks, but to become that domain subject matter expert to help guide and shepherd these applications towards leading to good positive outcomes for candidates and the workforce.
Matt Alder [00:17:31]:
So final question. Out of all of the applications of AI that you’ve seen or you’ve seen proposed, what was the most surprising one?
Rob McCargow [00:17:43]:
You know what, I think there’s so much attention that goes to the ones that sound sexier and more glamorous and the ones that get a lot of the media attention. For me, I think the ones that are actually the most eye opening are the ones on the first cut. Sound actually pretty damn boring. Let me give you one example of what we do here. So if you can imagine, we’ve got a 237,000 person organization, a multitude of skills and grades and expertise, and home locations, well as thousands of clients. So if you think about both sides of the supply and demand dynamic, the way that that is currently matched is a very sophisticated, albeit human led activity and this can lead to some quite, you know, negative outcomes. Why are we flying this individual from one part of the country to another on a daily basis? It’s not good for her productivity well being and it probably has an impact on our retention and attrition stats. Some of the some of the technologies such as machine learning allows you to do some amazing things with optimization. It can make sense of these patterns of data that is way beyond the human’s capability to understand. So some of these problems we often talk about having more permutations and there are atoms in the universe, so no wonder humans aren’t particularly good at that problem. So let the let the AI take away that allow your human resources professionals and recruitment professionals to focus on the relationships, the added value activities and make good sense and lead to far more productive employment outcomes.
Matt Alder [00:19:33]:
Rob, thank you very much for talking to me.
Rob McCargow [00:19:35]:
Been a pleasure, Matt. Thanks for having me on the show.
Matt Alder [00:19:37]:
My thanks to Rob McCargow. You can subscribe to this podcast in itunes or via your podcasting app of choice. The show also has its own dedicated app which you can find by searching for Recruiting Future in your App Store. If you’re a Spotify user, you can also find the show there. You can find all the past episodes@www.rfpodcast.com on that site. You can also subscribe to the mailing list and find out more about working with me. Thanks very much for listening. I’ll be back next week and I hope you’ll join me.