Transcript: A - Z of tech: Episode 1: A for Artificial Intelligence

Felicity

Hello and welcome to the very first episode of our brand new A to Z of tech podcast. I am Felicity, I’m in tech PR here at PwC.

Louise

And I’m Louise and I work in cyber security. Today, we’ll be looking at A for artificial intelligence. Which is of course just a small, casual, bitesize topic to start us off nicely with!

Felicity

Yeah a nice easy one to get into there! So we’re joined by our first guest in the studio, Rob McCargow, today who’s an AI director here at PwC. And then we’ve got a great interview Rob did with Tabitha Goldstaub, co-founder of UK AI start up CognitionX and chair of the government’s new AI council.

Louise

So without further ado, let’s welcome Rob to the studio, thank you so much for being here for episode one.

Rob

Thanks for having me on the show.

Felicity

I’ve always wondered about what the difference is between AI and automation, people are calling it robotic process automation, RPA, so perhaps you could tell us a little bit about that.

Rob

Indeed yeah.  And I think there’s constantly different versions of how to explain what AI is, I am not going to get too hung up on the technology too much. The best description I saw of AI and automation is that AI is around the thinking, and automation is around the doing. So, if there is manual tasks and repetitive tasks that can be automated, doesn’t require sophisticated technology, then RPA is proving very useful in those circumstances. But ones that are requiring an ability to discern incredible value from huge patterns of data - I don’t think intelligence per se is the right word to describe it as yet, because it is still fairly narrowly focused - but that is very much around augmentation, and I think, AI falls into that thinking bucket in that context.

Louise

I would be really interested to get your thoughts on the ethical and regulatory implications of this as well.

Rob

Sure, it’s an absolutely red hot topic. One thing that we’ve seen, some really positive progress in the UK over the last 12 months, is around how to address that very topic. Of the back of a number of government activities and initiatives, we’ve seen the creation of the AI sector deal, we’ve seen the House of Lords Select Committee report back on this, and we’ve seen government action. We have seen the creation of a Centre of Data Ethics and Innovation, we’ve got an AI council coming through as well, as well as a number of very exciting non-profits and another societal organisations. So, I think, the UK is in a very interesting position in how integrated we are in working with professionals from across lots of different disciplines to look at these multidimensional problems together, rather than working in siloes to try to address them.

My experience around AI suggests that we have a better chance of creating AI for all, that’s taken ethical considerations into account, if we have diverse group of people from different backgrounds contributing to the way it is designed, trained and deployed.

Felicity

So you’re a director in our AI team here at PwC, what on earth do you do every day?!

Rob

Well Felicity, it’s fairly wide ranging brief at the moment. You can imagine with AI, we have a huge body of work undertaking at the moment around deploying AI to ourselves to futureproof PwC, but also to work with our clients across all sectors to harness AI innovations, to transform their businesses as well.

There are a lot of other issues pertaining to AI, though that we get involved with, how does it affect the workforce in the future? How do we build trust and transparency into the technology and assure it? How do we consider some of the things, Louise, you work on around cyber security of AI systems and data protection. From a regulatory standpoint is quite a broad array of issues that affect AI, and we are trying to cover almost all of them.

Felicity

Nice easy brief then. So you are always going all around the world as well to loads of different conferences, what are some of the sort of things that people always come up to you at those conferences and say, ‘hey Rob, I am so worried about XYZ.’

Rob

Yeah, I think, depending on the conference, once you’ve moved beyond some of the longer-term issues, such as the singularity, science fiction, and Terminator, we tend to see some pretty big themes coming up in almost all of these panels and discussions as well. I think, the two main themes throughout the last 12 months have been, first of all, around the ethical considerations of AI – it’s a topic that’s received huge amounts of attention in the media and in business as well; and I think, the other topic is around the impact on the world of work, around jobs, education, and skills, and how AI can start having a significant impact on this in the future.

Louise

You’ve already alluded to the fact that AI is becoming something of a hot topic. If there was one thing that you could change about the current public AI debate at the moment, what would that be?

Rob

The topic tends to get quite polarised between either this very overtly positive marketing-led approach of how it suddenly transforms businesses by default and it acts in a very magic way to make life better for all of us.  Also though, there is a very opposite take on this, which is much more negative, much more dystopian, and much more jumped upon by certain groups when an AI breakthrough takes place. I don’t think this helps either.

I think, the best thing that could help us to move on to inform businesses, to inform the general public, is taking a calm sober course down the middle. It doesn’t sell press copy unfortunately, but to look at the very real opportunities this technology presents today, and increasingly in the next few years, but also looking at some of the potential risks that it brings to bear today and in the coming years as well, and maybe not get so hung up on the longer term picture, which has got some amazing research going into, but doesn’t have that bearing on life today.

Felicity

Mm, as we know it. Something I’ve heard you say in the past is the thing that really annoys you is Terminator pictures accompanying every story you ever see about anything pertaining to AI, is that something you want to eradicate if you were king for the day?

Rob

If I was king of AI for the day, I think, that will be first of my hit list. It is something that unfortunately is quite a lazy trope that’s attached to AI breakthroughs, AI equals robots equals Terminator and picture goes on the front page. The problem comes, I think, is around the way that the public now understands this technology. One of the big national or international conversations we now have to have is around personal data, around how we use data to train these systems to improve everyone’s lot, to drive economic growth, to drive innovations, to make life better for all of us and the planet. If we can’t responsibly articulate what this technology is, and we resort to some of these lazy memes, I think it erodes trust in the technology. There is great work happening with various institutions that we are very close to, to try to present this much more real, honest, and clear perspective of what AI means to people’s lives, to businesses going forward.

Increasingly less in the dependency of some of these sort of memes and tropes we see in the press.

Felicity

So something that you posted a while ago on Twitter, that went a bit mad, were the “Robot Rules” from your kids. So, how old are your kids Rob?

Rob

They are now, 9, and 7 and almost 3.

Felicity

So, your two eldest wrote four top tips for robots over the coming years, and you can find these on Twitter, we will link to it in the notes.

Rob

Exactly, over the last couple of years, I have been to so many events, where I have been inspired, I have been slightly frightened about the prospects of what’s coming down the track for the next generation. So I thought I would have a conversation with my young kids about this topic and it didn’t start off all that well.

I asked my son what he wanted to be when he grew up, and he said he wants to be a train driver to which I, kind of, dismissed that out of hand and said the robots will do that, to which he burst into tears and run off to his bedroom.

This then set in motion this really interesting set of activity to try to get me to help them in the future. First of all, I’ve sent them off to school to get a book on technology, he came back with a book on fax machines, which didn’t particularly fill me with confidence about the education system’s ability to prepare our kids. We also had an amazing experience by getting them to play MIT’s moral machine, which sets different hypothetical situations up for you to decide what a self-driving car should do. So my kids played this game, I think the main memory was, they preferred to save the lives of animals over humans, and they are very law abiding, so I think my parenting skills are coming through.

But all of this start to increasingly giving me some confidence that actually kids see these big megatrends in a different set of eyes to us adults with the baggage we’ve got. So we were about to present to parliament, a couple of years back, to explain to politicians what AI is, so I thought I’d enlist the help of my kids.  They gave me this set of rules by which to inform how the politicians should understand, what we need to do to make sure that AI and robots are good for humanity.

Felicity

There was some really great ones in there. Bad people shouldn’t build robots, number one, but also touching on something we talked about earlier, robots shouldn’t look like humans, and that was quite an interesting one.

 Rob

Well it’s an interesting one, there is a very particular concern in the AI community around this topic of, what is called anthropomorphism, the bestowing of very human characteristics upon technology. And if you’ve got amazing breakthroughs have happened this year, technology that’s able to impersonate humans on the telephone, coupled with increasingly life-like robots doing the rounds of various conferences and put the two together, there are very real implications for humanity around deceit and trust, and whether or not these things should be labelled as bots, if they are providing you with advice or interacting with the general public.

Louise

You mentioned during a briefing to the House of Lords on this particular topic. I wondered for those of us, myself included, who aren’t experts or even necessarily trained technologists, are there any good reading materials or sources that you could recommend?

Rob

Yeah, there is too much actually. I think there is some brilliant books that help to set the scene. I would recommend books like Richard Susskind’s Future of The Professions, Calum Chace and his Economic Singularity. If you want something further ahead, Max Tegmark’s Life 3.0 is a really interesting one as well.

Felicity

Yeah, I am reading that at the moment.

Rob

Yeah, it is a fascinating book.

To look at the shortest nearer term issues, there is great work done by whole range of bodies. I would start actually with some of the reports gone into government this year, The House of Lords select committee review on AI. I would also recommend organisation such as the AI Now Institute, the RSA and the Royal Society, have all produced fantastic reports, they are very accessible, in plain English, to explain what the state of the development of this tech is, the opportunities and the risks.

Louise

Brilliant, thank you.

Felicity

So that’s something that you spoke about with Tabitha Goldstaub recently when you caught up with her.

So Tabitha is co-founder of the UK AI start up, CognitionX, but also chair of this new AI council that we talked about. So perhaps let’s lead into that interview now.

Rob

First of all Tabitha, what do businesses need to know about the latest developments in AI in the UK.

Tabitha

What’s really exciting about artificial intelligence is the opportunity that it now presents for us to build products and services that both increase a business’ revenue, can even decrease their costs, and at the same time delight their customers. My top tip for those businesses, is to focus on building AI that is for their teams and the staff, rather than instead of. So if you are building AI for people rather than instead of people, you are usually going to have more success.

Rob

This raises some of these themes that I think are coming out in the media.  There is a number of negative themes that are coming through around the impact on the workforce, what do you think businesses need to think about when it comes to the impact on people specifically?

Tabitha

If businesses start by finding the things that annoy their workforce the most, and then work out a way to use AI to alleviate those issues, then I think they will be in for a brighter future rather than if they come up with problems to things that they just want to solve using “magic” for less known reasons.

Rob

There is a number of other topics that are coming through on a weekly basis around some of the downside impact of AI, whether it’s the issue of bias and discrimination, whether it’s leading to poor outcomes that are affecting different parts of society. Are these fears well founded, or does the press tend to run away with this and get carried away?

Tabitha

As long as we don’t think of them as fears, but more as concerns and issues that we need to focus on, then I think that it’s healthy. One of the most important things for us to do is recognise where and when an algorithm or an AI system is betraying attributes that we don’t want or we don’t expect. Actually sometimes the press is helpful, if it means that we’re more cognisant of this. Do I think that the press are running with somethings and displaying the facts in a more sensationalised way, yes; do I think that businesses should be more focused on the ethical considerations and think more carefully, yes.

All in all the opportunity is there for us to get this right, we just have to be more mindful of how AI is being built and who is building this AI.

Rob

That raises an interesting topic around the very real issues of AI leading to poor outcomes. In your experience, what are some of the practical steps that businesses can adopt to minimise some of the risk of harm or poor outcomes?

Tabitha

It really depends whether a company has decided to build an AI system themselves or if they are using off the shelf products, or they’ve partnered.  Technically, there are a lot of different ways that we should be looking at fairness, transparency, bias. The key is to interrogate whoever is working on those projects to ensure that they’ve thought of that. I also think that the quick win in many of these cases is to just make sure that the right people are in the room at the right steps. Rather than the engineering team, whether it’s your own or whether it’s an outsourced group, showing you an end result, actually have ethicists, English majors, people who are not from the engineering team from the other parts of the department involved in the process, they usually will be able to put their hands up and wave, ‘this isn’t okay’, a little bit quicker. As long as you empower them to do that, it’s a good way to save yourself a lot of headache.

Rob

Do you think we are doing enough in business to improve access to technology careers for women and other groups?

Tabitha

I think there is some really good initiatives in all businesses around trying to get more women into this space. Do I think we are doing enough, absolutely not, there is always more that can be done, we are so far from any level of equality in most engineering teams. There has to be, as Ginni Rometty calls it a ruthless focus on diversity and inclusion. A lot of diversity initiatives that don’t take into consideration, what then happens after somebody joins a company or after they join a group, and we need to think about actually including diverse voices, and making sure that they are empowered and able to make a change, rather than just box ticking. So, there is a lot more to be done.

Rob

What can our listeners do if they are really interested in learning more about AI, get involved in AI, where do they start, where do they go for advice and ability to get involved?

Tabitha

The good news is that there is a lot of information online, whether it’s in the traditional press or in niche publications. There is a lot of events that are happening in London. If you have the appetite, there are courses that you can go on. Ultimately though, above and beyond that, it’s about being passionate enough to seek out the groups within your own organisations that are already doing this. Most companies, I think the stats are well over 80% of big companies are already starting with this, some of them with small proof of concepts, some of them have already been rolling out AI for a while, but go and find them. Go and look through your intranets, and even if you are not technical, in fact most importantly if you are not technical, go and ask to be part of the working group, go and get involved, think up problems that you are facing every day that could be solved by AI, raise your hand and get more involved.

Rob

Fantastic, just one final question as well. For many of our listeners thinking about the impact, not just right now, but on the future generation, what should they be thinking about encouraging their kids to get involved with in terms of further education, to prepare them for this fairly uncertain future where AI could be really quite a profound force?

Tabitha

That’s a really good question and I think that actually the answer is in your question. It’s the uncertainty, it’s preparing young people for uncertainty is probably the key skill that we are going to need. Yuval Noah Harari’s book, Lessons for the 21st Century, has got a very good chapter that I would recommend listening to about young people. The thing that I really learned, and took away from it was, this isn’t about learning a specific skill, this isn’t about having knowledge in a specific area, because those two things are going to be readily available if they are not already online and in technologies like AI. So, what we need to get really good at is problem solving, critical thinking, being creative, thinking around the solution, empathy for other humans, and enjoying the challenge that comes with this uncertainty rather than being fearful of it.

Rob

Thank you very much for joining me today, Tabitha.

Tabitha

Thank you!

 

Felicity

A really interesting interview there Rob and the interviewee becoming the interviewer, you are going to be doing a few more of those for us over the coming episodes.

But perhaps looking to the year ahead, it’s 2019, what are some of the most exciting things on your radar that you think are going to happen this year?

Rob

We are going to start seeing a lot more progress in businesses adopting the tech. This last 12 months has been about quite low level investigation of use cases in the main, with some really promising breakthroughs.  We will see an expansion and an acceleration. I think we will also see a lot more diversification across industry sectors. The industry sector that I am particularly excited about is healthcare, and health industries, and the pharmaceutical sector. From a personal perspective, the reason I am most focused on that sector is, I very sadly lost my mother to motor neuron disease 15 years ago, and there’s amazing breakthroughs starting to happen, in particular in the UK around harnessing machine learning and other AI techniques to discover new drug compounds for rare diseases. If we can get behind that, if we can see progress in that sector, it makes the whole of society’s lot improved, it improves the economy, and I think that would be a win all round. So healthcare, in particular, is one to watch for 2019, but to be careful about badging things for AI for good, because it suggests everything else of AI is bad.

There are some AI bad use cases admittedly that unfortunately have come through in the last year. But there are genuine opportunities to apply AI towards the acceleration, towards achieving the Sustainable Development Goals, something we have been heavily involved with, whether it’s in assessing impact on climate change or deforestation, but also the convergence with other technologies, such as drones, for example, looking at the impact on humanitarian crises, etc. But I think, once you move along a level from that, I think it is just making life better, making us wealthier, making us healthier, and ensuring that some of our more intractable problems are taken away, we can focus on much higher value tasks and make everyone’s lives a lot happier.

Louise

Rob, thank you so much for coming to speak to us today and really nice to leave on a positive thought, looking forward to the future, and the exciting developments we might be seeing in this space.

Rob

Thanks for having me.

Felicity

 

And what’s your twitter handle to say everyone can find you, follow your musings?

Rob

You can find me on Twitter @RobMcCargow.

Felicity

Perfect.

So that’s it for episode one. Thank you so much for joining us. Coming up next, we will be hearing about B for blockchain, another nice and easy one to tackle. I know, a lot of people that I speak to certainly aren’t entirely sure what it is, so hopefully we can dispel a few myths.

Don’t forget to subscribe on whichever podcast app you use to get the latest episodes as soon as they are out.

Louis

Please do feel free to rate and review as well, so that other people can find the podcast. We would love to hear from you. If there are any particular areas you would be keen to find out a bit more about, so maybe you’ve got a burning question on a story you read this week, or maybe a new bit of technology that seems little bit baffling.

Felicity

You can find us on twitter @PwC_UK or I am on @FelicityMain. We are still battering Louise to get on Twitter, but I am sure that will happen soon.

See you soon!

Louise

Bye!

 

Follow us