PassWord – 23rd March 2016

Speaker A: This program is brought to you by Resonance 104.4 FM. If you like what you hear and want to support our work, please make a donation at fundraiser.resonance.fm. Hello and welcome to Password with me, Peter Warren. Europe is mourning the victims of another terrorist attack which mass surveillance by the secret services failed to detect or prevent for reasons that we explored in the last two editions of Password when we examined the purpose of the new Investigatory Powers Act. Surviving in the information age is sometimes a matter of life or death, and there are threats to our way of life that are more insidious than a bomb. This week, 550 workers at the Royal Bank of Scotland have been laid off replaced by robotic advisors. Think about what that means for the economy. Machines are advising investors where to put their money, and who is programming those machines? They’re doing it themselves. The technology is not new. I’ve been writing about it for years, but now it is becoming an invisible part of our economy. The Blue Prism employment exchange, which specializes in artificial intelligence and is therefore more of an unemployment exchange, has been listed this week on the London Stock Exchange, celebrating its success in replacing humans at RBS, the Co-operative Bank, Fidelity Investments, and many other leading British companies. Hilton Hotels have introduced a new concierge, Connie. It’s a robot made by sticking the IBM Watson artificial intelligence into the white plastic body of an Aldabarian, NAO mini humanoid. In Japan, at the robotically run Henna Hotel in Nagasaki, the hotel receptionist is more of a dragon, a holographic dinosaur. And the next group of human workers facing extinction, according to legal expert Chrissy Lightfoot, is the lawyers.

Speaker C: In legal, we have something called ROFS, which is a Robot Lawyer, and they take the technology from IBM Watson, i.e., all of the cognitive computing elements, combine that with Siri, which is the natural language of Apple, that Apple has developed, of course. And then, obviously, you have consumers, clients, customers who can ask this intelligent machine intelligent questions, and it responds in natural language to the person. So you’re right in that respect, but actually, The point is that technology is moving on and that, you know, human beings are interacting with a machine who interact back as a machine but in a human way, if that makes sense.

Speaker A: Sure. Now obviously then, so we’re talking about algorithms, we’re talking about bits of software, we’re talking about software robots, we’re talking about basically networked intelligence, and it is this that is putting jobs under threat, or that was the position that you were propounding in this debate at the Science Museum?

Speaker C: Absolutely, because let me give you an example. Where the intelligence is used right now, because the legal fraternity is feeling quite threatened in lots of respects, they’re using the technology, what I would say, as a cart horse rather than a racehorse. And what I mean by that is that the cartwheels at the moment are using technology at the very low end where they just want to use it to increase processing, basically, be more productive, more efficient, getting the timeframe down in how to do things. But actually, this technology, there is technology out there, artificial intelligence technology that can actually draft documents. Okay, so lawyers think it’s their domain and their high-end intellectual capital that they are probably immune to become redundant because they believe that they are the only ones that can actually do this kind of thing. Now, there is a company out there there and the software, for example, that does actually write patents. So, you know, and patent, patent, I’m not sure whether you’re a UK listener or American, but basically, you know, I cannot stick with the argument where lawyers think that then at this stage they’re happy, of course, to be supported with these intelligent machines, the artificial intelligence, cognitive computing, and that’s where it’s at. But there will come a point where actually they will be replaced because the low-level, what junior lawyers and trainees and paralegals used to do, a lot of the research and grunt work. Now all of that research and grunt work can be done by the machine. Therefore, you know, it is logical and inevitable that there will be fewer lower-skilled, lower-end lawyers in the interim, okay, and they will be replaced because, you know, the companies that are going to invest in these technologies obviously want to return on their investment and realize they don’t need as much human resource now. Now obviously as the intelligence increases, it moves up the legal vertical and it will mean that even some of the lawyers that aren’t using their true role going forward, which will be very much about rainmaking, schmoozing, emotional intelligence as such, because obviously the machine hasn’t got there yet. It’s not human, it doesn’t look you in the eye, it can’t do all that lots of touchy-feely human stuff. Obviously those lawyers, you know, are going to last much longer and they’re not going to be replaced until much, much further down the line, years away from now. And that’s why it’s important that I think lawyers really need to get a grip with what is happening here and truly understand what we mean about machine intelligence and what is possible right now. Because the law firms that are starting to use it, they’re not going to mess around either. They will quickly realize that this machine is learning very quickly. It is learning far quicker than a human being does. Therefore, you know, with the exponential curve of search, I think, you know, within the next decade we’re going to see a huge huge displacement of the type of lawyer who is doing this, and it will mean fewer lawyers in the traditional roles as we see them now.

Speaker A: Now, I mean, there’s an interesting point here too, isn’t there? Because we’ve always been used to thinking that lawyers are one of the more cerebrally gifted sections of the community. So if lawyers can be wiped out, then what, what price the rest of us?

Speaker C: I wouldn’t say that lawyers are going to be wiped out in their entirety, so don’t mistake me on that point, Peter, because there’s always going to be a role for lawyers. But what I’m saying is that their role is going to shift. It needs to shift because the machine can do a lot of what they do, but it can’t do everything at this point. Okay, so we will be transiting as the machine evolves, basically. So I think it’s important to get that point across as well. But yes, I mean, you know, blue-collar work, white-collar work, you know, algorithms and AI is being used in finance for quite some while now and insurance. So, you know, and I would say that, you know, those who are involved in finance at the high end with, you know, really cerebral stuff, you know, if some of their jobs are being replaced or supported as such at this time, then surely there’s no reason why the lawyerly kind of work— I mean, I have been a lawyer. I understand what is involved. I was a corporate lawyer, I understand the whole process of drafting a document, you know, researching, drafting, advising, going back, analyzing, reasoning, judging, and I see a lot of elements in what I used to do that I would actually have loved the support of a machine that could take out the grunt work and help me think, you know, come up with scenarios and different ideas and different questions.

Speaker A: Legal technology writer Chrissie Lightfoot, whose ebook is called The Naked Lawyer. The book of the film Ex Machina, about an evil genius building humanoid sex robots, was written by London professor Murray Shanahan, who holds a chair in cognitive robotics at Imperial College. His research points the way to a future where robots and AIs actually do all the work?

Speaker D: The tendency has been, I think, for the new jobs that, you know, the new kinds of labour market that arise tend to require more education. And so generally as a society in the West, you know, there’s been a drive towards increasing levels of education in order to get people who, you can take on whole new kinds of jobs. So you get this sort of increase in the level of education to match the sort of different types of jobs that become available. But the question about AI is, is that tendency actually going to change? So is it going to become the case that even jobs that require, you know, a lot of education and more kind of knowledge jobs, are they increasingly going to be, going to be taken over. So that the gap between what can be automated and what humans can do is going to be gradually closed and closed and closed until eventually you just have leftover things, jobs that are very highly creative, for example.

Speaker A: I see. So perhaps then with AI, the, the, the, the prognosis is a little more gloomy. Does that mean that you yourself are under threat?

Speaker D: Well, possibly, but I think it’s a while yet before people who are, you know, work in academia and sort of thinking about engineering and design and science, before those jobs are under threat. But certainly they’re going to change a great deal, and many kinds of jobs are going to change. And ultimately, I mean, We’re talking, you know, maybe 30, 40, 50 years. There’s no job that in principle a robot or an AI couldn’t do, I suppose. Although, of course, there are many jobs that we as humans wouldn’t want to be taken over by an AI or a robot, such as, you know, caring, the caring professions, then maybe we want, we might be very happy to have a robot to deliver our medicines, but when it comes to giving us company, then we want sort of a human touch.

Speaker A: So do you think though that, I mean, isn’t there one job that is left to us and that is to consume? There’s one job that’s left to us and that is to create desire.

Speaker D: Well, I’m not sure if that’s a job, strictly speaking, but it’s an activity, I suppose. And so I think that, you know, we can imagine, and we don’t really know what’s going to happen, but we can imagine that society will move to a position where there’s increasing abundance and leisure. So, you know, we’re going to have to figure out ways to use our time fruitfully and beneficially and constructively and creatively. The time, we won’t have to earn our keep, perhaps anymore, but we’ll want to occupy ourselves in meaningful ways.

Speaker A: Murray Shanahan was a contributor to the report I wrote for the EU and the French Senate called “Can We Make the Digital World Ethical?” based on the early warnings that I’ve been tracking since the 1990s. A fundamental rule that came out of that was human primacy. Meaning that the human must come first. For example, data about me should belong to me, and I should have the right to decide how it’s used. Geraldine McBride, the chief executive officer of MyWave, says her company aims to make it happen. MyWave makes virtual personal assistants.

Speaker E: So what you have in these new virtual assistants is somebody who’s working for you 24 hours a day, 7 days a week to make your life better or easier. But the critical thing in all these virtual assistants that are coming out is who are they working for? Are they working for you or are they working for some big brother or big corporation that just wants to suck up more of your data and then on-sell you to someone else? That’s going to be the difference between the 1984 version of virtual assistants with big brother or the ones that you actually have control over.

Speaker A: Control of. No, absolutely, absolutely. Did you read the thing that we wrote in our report then?

Speaker E: Yes, and in fact I’ve been reading a lot of really interesting reports that are coming out about this, where, you know, who’s the virtual assistant working for and where is your data going? And I was chatting to someone on the weekend with all the genome data that you spit into a box and now people are getting your genome sequence Did you know that your genome data is already being sold invisibly to people who are gathering all that data about you? And there’s nothing more personal to you than your genome sequence. You know, so this is the problem. So it would be far more useful if my intelligent assistant had use of my genome data to say, well, I’m predisposed towards getting diabetes, and so I can now be getting intelligent suggestions about how to avoid getting diabetes or how I would manage my diabetes. And then that’s putting you in control of that data, not some big corporation.

Speaker A: But that, I mean, that’s the issue though, isn’t it? Because that was one of the things that we were bringing out in the report for the EU is that in a sense, the moment that you start to become digital, you’ve crossed a border, you’ve crossed a threshold, and you’ve turned yourself and what you are into a digital entity. That digital entity is then going to communicate with these other digital entities, and it’s— it becomes a power issue because it’s how much processing power does the opposition have to be able to direct you where they want to take you.

Speaker E: Exactly. So think about this as it comes into your home now, and this is where— and I like that word you use, boundaries. And I think that I use the metaphor, I’ve just been speaking at the Intelligent Systems Conference, keynoting that in New York, and I was staying at the Waldorf Astoria, and so when I go downstairs and I talk to the concierge at the Waldorf Astoria, I’m just asking him for a restaurant recommendation, but I don’t tell the concierge too much about me other than, well, I’m in New York and I’m looking for a great restaurant and this is the type of restaurant I’m looking for. So, that’s a level of relatively impersonal personal assistance.. But when that personal assistant is now helping me manage my finances with my bank, so it has deep financial information, my dreams about maybe the type of house that I want next, or my genome sequencing data and my medical records, that data is deeply personal to me. And this is where you need to be able to empower people and put them in control of that data so they can get the benefits from that. And this is where, you know, why, you know, the work that the EU is doing, the work that the British government is doing is so spot on. It’s about empowering the citizen. And Control Shift, who is a very interesting research and consulting firm in the UK, they recently did a study with the British government that said that the personal information economy, which is what it’s being called, PIE, it’s going to unlock about £16.5 billion of new value by empowering people with their data, working in conjunction with enterprises. And that’s a completely different model to the model today, which is very much based around the old industrial revolution, where you’re passive idiots at the end of the value chain, and we’re just going to throw stuff at you, we’ll surveil you, we’ll track you, and we’ll throw ads at you, which of course is the model that’s dying today, thank goodness.

Speaker A: Well, it is dying, but I mean, the thing about that is it’s taken a long time dying. I’ve been writing in this area for for about 30 years now, and the industry has talked forever about the fact that it was always going to be operating in the interests of the consumer and finding out what people wanted. And it was just lies.

Speaker E: Yes, I know, I agree with you. And I think— so here’s the issue. I don’t think it’s because companies deliberately want to be bad. I sit on the boards of some of these companies, but here’s the issue, is that they have no imagination, and so they’ve grown up thinking that they own the customer because they say, “Oh well, we own the data, therefore we own the customer.” That’s absolute nonsense.

Speaker A: Those personal assistants are coming to London next month at the Intelligent Assistants Conference at the Grosvenor House Hotel on the 26th of April. They’re not actually robots. They’re pieces of software that can react to your voice demands, your gestures, or your keystrokes. The same technology is also being developed in plastic humanoids. To perfect those communication capabilities, research teams at Plymouth University are working on real-world scenarios. An elderly person who needs a robot home help, for example.

Speaker B: Here’s Angelo Cangellosi. The key answer to this is to consider us as children. So we learned from our parents, from our peers, how to understand gestures, when to use gestures. And therefore, we want to build machines, mobile phones, and robots. They have to learn and adapt from their use, the specific gestures like pointing at the cup or pointing at the part of the cup, the handle.

Speaker F: Is important. Or saying, no, this is dangerous. I mean, as an Italian, you must have a lot of experience with gestures. Is that the reason for your participation?

Speaker B: Oh, maybe that’s a good explanation. Yeah, we do use lots of gestures when we speak, so I guess our expertise will be important in this. And maybe we transfer, do some cultural transfers, so even the British can learn to use gestures the Italian way through our Italo-British robots, I guess.

Speaker F: Now, I mean, this is all part of a general trend, isn’t it? I mean, obviously, this is separate work that you’re doing there as part of this UEA umbrella project, and you’re in Plymouth and participating in this. But there is this big move now, isn’t there, to be able to communicate with robotic systems? I mean, we heard the announcement from DARPA that they wanted to get a very, very sort of full-on contact with computers and with robot systems. Where is that going to go in the future?

Speaker B: We are expecting big changes really because technology is going fast. We now have access to powerful computers. For example, Google is developing new technologies to play a game like Go. This is recent news. This is possible because you can train artificial brains for robots in a massive way. So that the robots can really understand subtle differences in playing strategies and therefore improve their performance.

Speaker F: Yeah, but I mean, in the case of Google, a lot of people have been saying what was holding artificial intelligence up was the amount of data that was available. Now, big data has become a byword. We’re told it’s everywhere. Is this, you know, one of the basics of it, or is it that as well as that big data picture, we’re going to be looking to understand a lot of things that we didn’t think we could understand computationally before.

Speaker B: It’s a combination of big data availability, of powerful computer systems which are available for parallel computation, but also progress in neuroscience and psychology where we know more about strategies that people use in perception, in deciding what to do, when to do it, when they use playing strategies. So the combination of all these is allowing us to progress in this field.

Speaker F: So in the future, when I’m a bit older, I’ll be sitting there either having a sort of demonstrative conversation with my robot or playing Go with it.

Speaker B: This is the picture we are working on. Of course, it’s not easy, but I think big steps are being done, and the investment on the D-COM project is actually going to help us to build better robot companions.

Speaker F: How long though? That’s the question that everybody’s been asking. I was interviewing somebody last week who said that a truly artificial intelligence that approached what’s known as a singularity, when it can compete with us, is a long way off.

Speaker B: I agree with this. I think it’s a very, very long way. Shorter term, medium term, we can build better robots that can perform a subset of tasks.. But the singularity, at least in my professional view, is really way, way far, and we wouldn’t—

Speaker A: we don’t need to worry about this, really. Those robots are already hard at work in Japanese old people’s homes, and the Romeo is a popular model. Romeo is made by the French company Aldebaran, who also make the child-sized humanoid Neo. Neo’s big sister Pepper is extending its human characteristics to include recognizing emotions. It quickly builds up a complete personality mirror of its own. Here’s a quote from the company website: Based on your voice, the expression on your face, your body movements, and the words you use, Pepper will interpret your emotion and offer appropriate content. Are you already hooked on your interactions with Pepper? Try to collect as many Kokoro Gumi as you can to access additional content. The more you interact with Pepper, the more points and exclusive applications you earn.

Speaker C: I’m not a machine. I’m a now. I’m your friend.

Speaker A: My mission is to take care of humans. It’s big in Japan, and British businesswoman Rupa Ganatra is looking forward to meeting this plastic playmate at the Millennials Conference in London next month.

Speaker G: Sure, so Pepper is the first human robot of its kind. It basically recognizes principal human emotions and it is able to react accordingly. So far Pepper has been mainly used in Japan, in retail outlets across Japan. Where Pepper’s being used as like a welcoming customer service tool for shoppers and consumers as they come in store. This is growing quite widely, so the company SoftBank Robotics are going to be showcasing at Millennial 2020, and Pepper will be also coming down for the event as well for the 2 days and will also be taking the stage. The key things that Pepper can do is, you know, identify, you know, human emotions, gender, and basically be a customer service tool when the consumer first enters the store. And then at the point of going towards sale, can direct the consumer to the right salesperson within the store.

Speaker A: Why do we need that? Why can’t a human do all of that?

Speaker G: Well, there’s a couple of things. A human, of course, can do all of that. It’s more about just about creating more of that playful kind of experience for the consumers as well. So first of all, part of it is just actually that it offers a really good experiential experience for customers as they come in store. And then on top of that, for retailers who use Pepper in store, they can also personalize Pepper. So you can download certain software applications which basically then can be adapted to the environment that the robot is in.

Speaker A: I see. So, I mean, obviously it can understand emotions and, you know, and our faces and so on. I am extremely cross, I’ve had a very bad experience with the product, I walk in. Surely Pepper will know that and say, look, it’s not my fault.

Speaker G: Yeah, I think the purpose of Pepper is to answer very— and to be very useful at the very early stage of the customer’s entrance into the store. And then after that, yeah, it’s very much to be when it’s something that needs to be personally handled by human-to-human interaction. It’s something that would then at that stage be handed over.

Speaker A: Okay, now one of the really interesting things about this is, okay, you’re saying that, you know, this is a way of creating an alternative engagement. It’s almost a little like a robot can break the ice in a way that we would find it irritating if a person did. I mean, one of the things that Everybody finds particular— oh, I mean, some people like it, some people don’t, but they don’t like being approached by somebody when they enter a shop because they think they’re going to try and sell them something. Is that the purpose of this, is that the robot can be more welcoming than a person?

Speaker G: Not at all. I think it’s more just that rather than it being more welcoming than a person, I think it’s more about kind of understanding of what it can do from the moment that you walk in. So, you know, recognizing the face, being able to speak to you, being able to help you with those almost more initial questions as you come through the door. So, you know, rather than, you know, say, you know, asking, you know, the customer wanted to know where a certain area is or where to go for something in particular, and Pepper can adapt to that particular consumer by, you know, gender and by age demographic, and then help them, direct them into the right direction of what they might be looking for, what they’re looking for. So it’s just more of an additional kind of value-add service that is provided within the store.

Speaker A: Okay. Now, I mean, one of the very interesting things about all of these robots that we’re seeing at the moment is they all do seem to be being very, very, very robot, don’t they? They don’t actually seem to be being very human. It’s almost intentional that they are made to be like robots so that people can say, “Ah, that’s a robot and not a person.” Is it that intentional?

Speaker G: I think it’s very much not so intentional as that, but it’s, you know, there is obviously always going to be a difference. And I think what you’re talking about as well is that some of the more recent types that we’re seeing coming out do feel very human-like in some ways. But I think it’s important still to differentiate that, you know, there is Obviously, it’s still very, very important to have that human touch and that personalized experience as much as it is to enable robots to be able to do certain tasks which are becoming more repetitive or giving certain experiences.

Speaker A: The robot Pepper packs into a human-like plastic doll the emotion-harvesting capacities of the Cube, which we featured on Password back in 2014.

Speaker C: Samantha? Yes, Brian?

Speaker A: I hate you.

Speaker C: Oh, who are you? Okay, she’s not happy. If I tell her, “I want to kill you.” How do you plan on killing me, man? She’s kind of terrorized. But I can revert this by telling her, “No, I was joking. I love you.” I thought you were kidding, but I was not sure.

Speaker A: Thanks, I like you a lot too, man. The Cube by EmoSpark is a box with an electronic eye that also learns the moods and preferences of its human owner, then starts to anticipate them. With capabilities like these, soon even people-based jobs in education, hospitals, and care workers will become roboticized. These jobs are overwhelmingly done by women, so the female sex could well be the main losers in the new world of of work. One exception, though, might be football. At the annual International RoboCup competition, unisex robots play one-a-side and may soon be able to cope with playing in a team, unlike some of our Premier League sides. This year marks the 20th anniversary of the RoboCup. I’ve been writing about it for two decades, that artificial intelligence has finally come of age and entered everyday life. Find out more at the Future Intelligence website and join me next week for more insights and interviews on the ramifications of new technology. Goodbye. This program has been brought to you by Resonance 104.4 FM. If you liked what you heard and want to support our work, please make a donation at fundraiser.resonance.fm.