PassW0rd – 10th January 2018

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.

Speaker B: Hello and welcome to Password with me, Jane Wyatt, in Resonance FM’s monthly exploration of technology in society. Today we ask, can we trust the machines? We’re talking about breaking into microchips and building blockchain, natural language processing, Intelligent assistants and intelligence agencies. The machines are in our homes, our workplaces, our healthcare, and wherever we go. We’re training them to know our likes and dislikes, voices, faces, and personal details. Perhaps we’re building them quite slowly, bit by bit, with each new piece of data, but the process is well underway. Artificial intelligence and machine learning can even create immersive This is ‘Cor Sonore’ or ‘The World is Sound’, an interactive installation. Its 4-month run at the Rubin Museum in New York City has just come to an end. It’s unique in that all the visitors and the museum itself were the art. By touching, looking, lying down, walking up the stairs and every kind of interaction, the public themselves created the sound. It was intended to be felt not only through the ears, but also inside every part of the body. Artists including Laetitia Sonnami and Éliane Radigue used drone sounds, synthesizers, and swinging speakers, as well as machine learning, to create the piece. Laetitia Sonnami says this process of letting the machines take part is unpredictable, like riding a bull. Artist Joachim Montessuis explains: I like to think of modal music as a skeleton on which we hang a melody. With other music, music without drone sounds, it’s another story. It’s much more mental, much more intellectual. With the drone sounds, you come out of the intellectual because they make you able to feel the sound. There’s a permanence, the body can recognize it. We’re made up of 95% water. There’s a resonance. Here the machines or algorithms are making people’s bodies and the architecture of the venue into instruments on which they play their own music. It’s beautiful. But man-machine creative collaborations can have a dark side. 200 years ago, a book was published on New Year’s Day that wrapped up all our fears about mechanical monsters: Frankenstein. Mary Shelley wrote her story on holiday with her future husband Percy, the poet, Lord Byron, and their friends from what we now call the Romantic Movement. It was a wet June on the shore of Lake Geneva, so wet that it became known as the year with no summer. So they amused themselves with a writing contest: who could make the best ghost story? The winner was only 21 when it was published, but her dark tale continues to fascinate. In the 21st century, it’s being used as a way for philosophers, technologists, and artists to confront its central theme: when we make a machine in our image, will it turn against us? Bob Beard is the project manager for the Frankenstein 200 project at Arizona State University. I asked him why this Gothic horror story is still relevant today.

Speaker C: Frankenstein isn’t really even a science fiction story anymore. It’s a story that high school students around the world enact every time they compete in iGEM, for example. Its message of scientific creativity and responsibility is something that we’re wrestling with on many technological fronts, from artificial intelligence to synthetic biology. And despite being 200 years old, it still strikes right at the heart of that dilemma. What are our responsibilities to the things that we create? And how do we define life? And how do we define ourselves as human beings?

Speaker B: To what extent, Bob, does artificial intelligence represent a new monster?

Speaker C: I think what’s most interesting about AI right now is not the monster story, which we’re all familiar with through narratives like the Terminator movies. What’s really interesting is the deeper lesson. In Frankenstein, it’s about how we respond to strange and alien life. The intelligence that we’re creating in machines right now is alien, and we’re not very good at dealing with that sort of strangeness. We tend to pack it up into black boxes and we instrumentalize it without grappling with the deeper implications of systems and machines that are already, in some simple ways, alive.

Speaker B: And what will people experience when they come to the Frankenstein 200 exhibition or join it online at your website?

Speaker C: Frankenstein 200 is a free, interactive, multiplatform experience for kids and families that’s designed to inspire engagement with emerging technologies. Frankenstein is a pervasive myth experienced through a web of stories and interactions from movies and TV to comics and toys and breakfast cereals, and our project really follows that. Audiences can investigate a mysterious digital narrative. They can go to a local museum or science center or library and do a hands-on activity. Or they can download a set of instructions for an at-home DIY making activity, all of which encourage people to be playful with the story of Frankenstein. After all, the story wouldn’t have endured if it weren’t fun. But working through all these elements, and in a very Frankensteinian way, assembling all of the pieces, moving from the digital to the physical, audiences can have an experience that is instructive but also accessible and personally meaningful.

Speaker B: Hmm, I’m not sure I’d want my children to use a do-it-yourself at-home kit to make a mechanical man. Still, if it inspires people to use technology to express themselves creatively, then it’s probably just the next step up from Lego and Meccano. Frankenstein 200 is at Arizona State University and the Bakken Museum in Minneapolis. But you don’t have to go there to join the debate about mankind and machines. The University of Northumbria in Newcastle is calling for papers for a conference in June 2018, and there’s a parallel conference on the shore of Lake Geneva, close to the spot where the Frankenstein legend was created. And just a few kilometres away, in the small Swiss town of Zug, a new technological creature is being nurtured: blockchain. Zug has declared itself the blockchain capital of the world, and a group of technologists and bankers held a meetup there this week. Under the banner of the Crypto Valley Association. The CVA is a far cry, more of a yodel away from those young romantic poets. Already global corporations including UBS and PwC have set up headquarters in Zug. Now, no one’s saying that blockchain is a monster. Well, apart from a few environmentalists, that is. Eric Holthaus, an American meteorologist, claims that in a few years Bitcoin mining and blockchain computing will use as much energy as the whole country of Denmark, contributing to global warming and extreme climate change. So maybe blockchain is another destructive machine rising up from the wet, windy shores of Lake Geneva. British scientist Dan Williams of KingTech takes a more measured approach.

Speaker D: Blockchain can be thought of as a ledger of information, and in the case of Bitcoin, it has financial data. This ledger contains every record of change to the item being tracked since its inception. These changes are added to the previous information in blocks which relate to the previous blocks, hence the term blockchain. The process of writing and validating these new blocks, called mining, is highly computationally intensive. Each block must be encoded using a process called cryptographic hashing, which is repeated many times over to fulfill the technical requirements of Bitcoin, and that makes it highly computational and the reason why so much power is used. It’s also the reason why blockchain is so secure. Of course, that’s its core benefit in that it makes blocks irreversible. Because Bitcoin rewards mining with Bitcoins, a computational arms race as such has been started with more and more power resulting in more advanced mining. It’s almost a catch-22 situation. Key to note here though is that this is only valid for Bitcoin and not necessarily all types of blockchain.

Speaker B: So what’s the extent of the problem?

Speaker D: In truth, no one can say how much energy has and is being used by blockchain technologies such as Bitcoin. Some estimates have put Bitcoin’s energy consumption at around 4.4 terawatt-hours per year, which is equivalent to a small city, and some have gone as high as saying it’s 35 terawatt-hours per year, which is roughly equivalent to the annual energy consumption of Denmark. It really requires academic peer-reviewed research to find out what the extent is now and in the future. It may change. However, in any case, the power consumption is significant, especially if this style of Bitcoin blockchain continues. It’s definitely worth further investigation.

Speaker B: And what will be the consequence if they don’t adapt?

Speaker D: I think it’s certain that Bitcoin will increase its energy consumption over time going forward. To what extent really needs to be determined, and I think academic rigor and peer review play a part in this. I think also this isn’t just an issue for Bitcoin, but it really relates to all digital services, especially those that come from large data centers. For example, cloud computing and media streaming. I’ve actually quantified a lot of cloud computing services and their environmental impact and seen it can be both better and worse for sustainability depending on its application and operation. I also think that questions revolving around the energy of the current finance system need to be explored scientifically, and that’s in order to see if in the long run a transition to a digital currency would in fact be better overall. But right now we really can’t say. Finally, I think sensational claims might focus on the today, but many ICT organizations are transitioning to a more sustainable world and way of doing things, and I think there really is a reason to be optimistic in the long run.

Speaker B: And it has many attractions. It offers transparency with anonymity. Every time there’s a new transaction, it appears in the blockchain online ledger. Swiss banks, with their numbered accounts, have always made anonymity a priority, so this is a logical progression for them. Where transactions have to be exact, like the price of a unit of cryptocurrency such as Bitcoin, it’s vital to keep the ledger totally accurate. Blockchain guarantees this accuracy. But there are many other types of personal data. They’re less exact, but just as important to the individual. Professor Irene Ng, of the University of Warwick is working on the HATs—individual online repositories for our personal data. At a conference in Bergen, Norway last week, she spelled out the advantages of using HATs for safeguarding personal data, as she explained to Password’s Peter Warren in an earlier interview.

Speaker E: It’s really about data, about people, individuals. You’ve got all these big cloud architecture, cloud computing facilities, And they are really out to help a lot of companies, organizations, and they help them with the supply chains and information systems. That’s great. But a lot of the data that’s about us, our finance, our health records, and everything, they’re all sitting in different vertical silos. And you think, well, if it came back to me, if I had a powerful information system that could search through my history, match what I have with what I was supposed to buy, If I could use this data that everybody seems to be holding off me better, and if I had a computer or information system that could help me do that, then surely this will help me as an individual. And that’s basically what the HAT and ROMPO is about.

Speaker F: Okay, so if you’re going to do that, where is this system going to be culling that information from? Is it going to be taking it from your computer? Is it going to be going out onto the web and doing it?

Speaker E: What you do when you sign on to a HAT is you sign on to your own little microservice container up in the cloud. This is your server. This is pete.hubofallthings.net. It’s your server. But we’ve built a backend to it that allows you to grab your data through APIs. Now a lot of these APIs is already being shared by different companies. Like if you sign on something on the internet today, you could use Google to sign on. And then suddenly you’re passing your email and information over to another company. Well, if they’re passing my data between themselves, why not pass it back to me? So there’s already a way to share that data. So just, I’ll take all my Facebook data back, my calendar data back, my email data, just give it to me.

Speaker F: So does that mean that we’ll have greater visibility on all of this information that’s out there? I mean, what’s happened is that we’ve become what people are saying is the oil in the information age. Does that mean that we’ll see our own oil? Exactly.

Speaker E: And not only do you see your own oil, we can make better use of it because these people, the organizations out there that have oil data about you, they really can’t share because it’s against Privacy Act. Now, unless they scrub it and anonymize it, they can’t really share it. And so they can’t give us really good services either because of this, of the privacy problem. And you get market failure. If you gave it back to me, I can share my own data with myself. I can mix and match the location, the people, the event. If I had a powerful enough information system, make it searchable, indexable, clusterable, I’ll use it and maybe I’ll make better purchase decisions. 65% of GDP are purchases from within the home. If we had better information, data on what we have and what we match to what we buy, buy, that’s a huge impact on GDP.

Speaker F: Does that mean that I can actually start to maybe make some value out of my own data, not just in terms of those better purchase decisions, but say, look, I’ve improved my data, it’s now better for me to buy things with, and I’ll sell that back to you?

Speaker E: Exactly, yes, that’s what the hack does. Under Freedom of Information Act or data protection, you can go get your data back from Tesco, you can sell it to Facebook. Now it’s really disruptive You can imagine, which is what I don’t want to say too much about, but it is possible. If you had a way to take it back, you can sell it. And it’s worth something. We as individuals, but the only problem is, as individuals, we are such high in variety. We’re so heterogeneous. But if you had a platform that could let you retain the uniqueness of your value of your data and the content, but yet had a standardized way of having protocols the way it’s exchanged, the way the data is moving in and out, then We as a collective become very powerful even as buyers of purchases, buyers of data services.

Speaker F: It’s not going to make you very popular though with some of the large data companies, is it?

Speaker E: Oh, I think we will be incredibly popular because one of the biggest problems with the data companies, and I’m not talking about the Googles of the world, I’m talking about the marketing, advertising marketing of the world, which is a $30 billion industry, is on the race to the bottom. I remember when I first got on the internet back in 1991, the click-through rate of a banner was 98% because everybody was curious what a banner was doing. It brought you to another website. Today, a click-through rate on a banner is 0.1%. We have ad blockers, as many as we have ads. It’s not reaching out. It’s not getting the right conversations with us as individuals. And so instead of trying to pound us with things and guessing what we are, if we had our own data, like if I had my shopping basket, you could tell me which supermarket around me is the cheapest today.

Speaker F: I mean, what you’re sort of saying is that we’re coming up with a problem to a solution, really. Exactly.

Speaker E: And if you think about— just imagine if I had my cupboard inventory. You are also going beyond the financial economy. You can say, for example, I need a cup of— I mean, back in those days, we need a cup of sugar, for example. You could just Which is, you know, if we had our kitchen inventories within, with our neighbors, we could share things that would be perishable tomorrow, or we don’t mind sharing. So we create a different kind of economy with data sharing that could have a huge impact on the financial economy, but it also has a huge impact on the way society shares and engages.

Speaker F: As a final point then, what you seem to be saying is that the internet has actually ossified to an extent And now the next phase, the next iteration of the internet, is that it’s going to become very useful again. The, the fact that it’s just essentially a collection of interlinked brochures showing various bits of information, it’s not active enough, it’s not engaging enough.

Speaker E: And more importantly, it is also representative of the inhabitants within it. They were not democratic enough. The reason I say that is because Tesco has a server, Amazon has a server. Where are our individual presence on the internet? We’re like cattle, can be harvested, we can be milked. We need a presence because when we have a presence and we’re collectively in a standardized, we have a presence, collectively we have a voice. We can do more together on the internet. We’ll make the internet better. And the reason why the internet isn’t as good as it is today is because a huge number of people who inhabit it are treated like cattle.

Speaker F: Just one final point though, and there is one very important issue that we haven’t touched on, and that is security. If you’ve got all of your information stuck in one place, given there are millions and millions of cyber crimes a year, the one thing that we don’t seem to be good at is protecting our own information. Is that not a problem?

Speaker E: So the way we’ve deployed the HAT is that every person has a microservice to container. This is why the hack is so unique. If you hack Yahoo, you will get a few million, hundreds of million accounts. If you hack a hat, you hack a hat. It’s a seller.

Speaker F: So you only get, you only get one account? Correct.

Speaker E: And so your best security measure is try to be as unimportant as possible.

Speaker F: Well, it’s wonderful, isn’t it? Because obviously the value, again, the value to the criminals is the oil. It’s the collective agglomeration of information. Exactly. And so if you take that away, then they basically have to make the efforts to hack you individually, which won’t give them a great return. Exactly.

Speaker B: Still, there are some aspects of personal data that are truly intimate. Concerning healthcare, for example, how can we trust an array of different apps with these statistics and preferences? And yet, if we don’t trust them, how can the healthcare professionals gain potentially life-saving knowledge from the pools of data that exist? One company that’s trying to take an ethical approach to this dilemma is Clue, the women’s health tracking app. Clue aims to give women insights into their menstruation, fertility, and contraception needs directly in their smartphones. Password reporter Charlotte Roundtree spoke to Ida Tin, Clue’s founder, at a rather noisy conference venue. So I think if we, for instance, bombarded people with ads that were very targeted because we know a lot about the health story. Like, I would find that inappropriate. I think you can ask people if they would like to have recommendations for services and products that might be truly helpful, right? If you have something going on, you might actually want to purchase whatever service, but that has to be kind of like on consent base, I would say. Have companies shown an interest in buying the sort of the data that you produce from Clue? People are very interested in having the contact to the users. I think that’s the biggest thing, because a lot of people that sell products to women, like the pill manufacturers, or people selling tampons, or whatever, they don’t necessarily know so much about how their users use their products. But again, we are very conservative. We have not made any partnerships with these. I mean, we’re conservative when it comes to protecting our users’ interests. We’re not conservative when it comes to thinking about being progressive.

Speaker E: Aggressive voice in female health.

Speaker B: But yeah, people are very interested, but we are, you know, we have not been so interested in that so far. Clues Ida Tinn talking to Charlotte Rountree. This is Password on Resonance FM with me, Jane Wyatt, and after this you can hear A World in London with DJ Ritu. The idea of taking control of our own health and personal data is very appealing. But here again, we’re forced to put our trust in an app controlled by Clue, who, as we heard, are even open to the idea of sharing details of our menstruation cycles with commercial companies. Even more worrying is the latest cybersecurity scare involving vulnerabilities nicknamed Meltdown and Spectre. They affect the very microchips that run most of our computers. Is nothing safe? Password editor Peter Warren, who chairs the Cybersecurity Research Institute, Told me more about Meltdown and Spectre.

Speaker F: What do they do? Well, basically there was a weakness that’s been around in Intel chips for some 20 years. That was the Meltdown weakness. The Spectre system actually can attack many more chips. So it can attack ARM chips, Intel chips, and AMD chips, which basically covers off most of the chips in the world. And the weakness is one in the way that people use computer chips. One of the things that everybody wants whenever they have anything to do with technology is speed. They have this insatiable desire for speed. So one of the things that the computer manufacturers, chip manufacturers, came up with was a system of trying to make the computers work faster than they actually should. And it was known as speculative execution. So if the computer chip thought it knew what you were doing, it would get ahead of itself, and so— or what you were doing, what you wanted— and it would go down a particular path. If it found it’d gone down the wrong path because that wasn’t what was required, then it just tracked back to the point that it was wrong and then went the other way. Now what’s happened is that while it’s doing that it leaves memory. It uses cache, i.e., it uses memory resources to hold information on what it’s done. And that information on what it’s done can be accessed when it’s in that cache state. When the computer realizes the wrong or the error that it’s made, it deletes that data. Now, why is this bad? Well, one of the problems that has occurred is that this means that data that is meant to be very, very highly privileged data can be left open and can be accessed by lots of other computer programs. So hitherto it was thought that it wasn’t possible for one computer program to be able to get access to data from another computer program. This means that this can happen. The other thing that the vulnerabilities allow is for you to get into the kernel, i.e., the heart of the chip. That is the inner sanctum of the computer itself. That has come as quite a shock to those people who are working in technology.

Speaker B: So what do we need to do about Meltdown and Spectre?

Speaker F: Well, to be perfectly honest, the majority of us should do nothing. Apart from doing what you should always do, which is install the upgrades and the updates from the computer manufacturers. The computer manufacturers are rushing out patches now, even as we speak, that are addressing this issue, and hopefully that will make sure that it is not going to be such an issue for everyone.

Speaker B: Since this vulnerability has not openly being used to break into anybody’s system, is it possible that it might have been either deliberately put there or secretly used by intelligence agencies, by state actors on the quiet?

Speaker F: It’s been speculated by many people that intelligence agencies would have been aware of this, that they would have been exploiting it. The NSA have denied completely and utterly that they, they have had any awareness and, and they had exploited it. To an extent, they have to say that. They wouldn’t say if they had been doing that. Yes, it could have been exploited. If it were exploited, it would be quite a difficult thing to do, so you would have to pick your targets very, very carefully. Yes, you could get some very, very valuable information, But you would have to be very, very good at programming, because one of the things that you would do if you wanted to take advantage of this is you’d have to be able to get into the chip, and you’d have to be able to do targeted sweeps for data that was trying to obtain the data before it could be deleted. So you’re in a sort of arms race against the system. That said, if you were good at doing it You would have been able, theoretically, to have obtained things like passwords to the overall system. That would have been very valuable to you. That’s the sort of thing that would be valuable to an intelligence agency if it were targeting a particular individual. But that we’ll never know.

Speaker B: If the microchips that form the inner kernel of our computers are at risk, surely we need to be able to lock down whole systems. One of several cybersecurity products that promises to do just that is Balabit, made in Hungary. Product manager Peter Gyonaszibi explained to Peter Warren how it works.

Speaker H: Our solution is to use a number of different algorithms that all look at the data from different angles. So it might be possible to trick one algorithm, but it’s quite difficult to trick 15 different algorithms at the same time while still managing to do what the attacker wants to do, steal the data or do something malicious within the system. Now, I think one of the things where I suppose in a sense you have an advantage is you know that an attacker— there was some research that came out very recently which said that attackers by and large, this is from the Office of National Statistics in the UK, they are after money. So if you know that there are certain things that they are after, i.e., the details that will allow them to get into an account to steal passwords or— you know, there are particular targets. That must make the job a little easier. To some extent it does, but we know usually what they are after. We know where the crown jewels are. What are the things that they want to get, be that information or be that performing a fraudulent transaction. But on the other hand, we have no idea about how they will try to achieve that goal. Security is inherently a cat-and-mouse game where the defenders, the security people, usually have to play catch-up with the attackers. Who are finding new and new vulnerabilities and attack methods to infiltrate these systems. This is where the expression zero-day vulnerability comes from, because it’s something that has just been discovered or, well, just been used for the first time. So using machine learning or artificial intelligence can be a solution to, to these problems. So instead of the traditional security measures very often try to come up with rules that describe possible attacks, and describe what is allowed and what is not. And in a situation when attackers are coming up with new and new ways of infiltrating the system, that does not really scale and it requires a lot of effort to maintain those rule sets. Instead of that, our approach is to have a system that learns what’s normal and tries to find anomalies, unusual events as compared to that baseline. That allows us to catch those, you know, those famous unknown unknowns. Things that we didn’t even know we should be looking for. And it is not only us who are doing this, the entire emerging field of user behavior analysis. There are some things that we do that I believe are quite unique, but the entire approach is indeed done by some other vendors as well.

Speaker B: Well, that may be good for security, but for the employees whose every keystroke is being scrutinized, assessed, and stored, It shows a lack of trust in the humans and a strong reliance on artificial intelligence. Ballabit’s 15 different algorithms can spot an intrusion because they recognize the company’s own workers from their familiar characteristics and patterns of behavior. But what if a genuine worker was tired or just not concentrating? Would a small mistake or a break from routine become a damning piece of evidence according to the software? It’s worth finding out whether your employer uses a system like this. Or, if you are an employer yourself, to explore the workings and administrator rights of your cybersecurity system. All of us can now keep track of our own consumer habits thanks to virtual personal shoppers called intelligent assistants. Geraldine McBride of the Australian startup MyWave invented FRANK, an app that facilitates easy price comparison and switching between energy suppliers. She believes the time is coming when we all need intelligent assistants.

Speaker I: The average human being can cope with about 9 apps, and it’s actually a scientifically proven fact. So, if you actually look at it, there’s a lot of app fatigue out there with consumers, and most of their personal data is siloed over all these apps, and it still requires me to go and do an awful lot of work. So, you know, like even going to a comparison website or going to any website, I still have to keep entering my data multiple times. It never remembers me. I have to remember all these passwords., and it’s just plain annoying. It’s too much friction. And what we’re seeing now is this new neural network of sensors, of intelligent agents that are coming across the world that are actually just making things happen automatically for people. So, you know, you say to your intelligent agent, you know, “Book me a train, get me that air ticket, find me the nearest restaurant.” You know, different intelligent agents are doing different things. And so, it’s creating frictionless commerce. And if you look at the intelligent agent, if you’ve got really smart ones that that are working for you. They’re almost like the new brains or the new neural networks that are coming out across the world. They’re replacing the browser, they’re replacing search, and they’re replacing apps. And they’re coming to a big brand near you. And so that’s sort of the high level. And then what I can do is I can talk to you about the different types of intelligent agents, because not all intelligent agents are created equal.

Speaker H: But also, you’re going to have to get one that actually you’re sure it’s loyal.

Speaker I: Looking after your interests. You’ve got to make sure who is your intelligent agent working for. So let’s think about those levels of intelligent agents. There’s a lot happening now with Facebook’s M, Google’s intelligent agent, Amazon has Echo. The problem is that those intelligent agents aren’t actually working for you. They are taking a lot of that data and you don’t know where any of that data is going about you. So those intelligent agents are actually working on behalf of Facebook and on behalf of Google in order to turn you into a product. So they can sell you more things. In fact, in the end it drives more advertising. So, that’s one form of intelligent agent. The other reason why so many brands were at the Intelligent Assistant Conference is that they are terrified of being disintermediated. So, imagine if you were a big airline or imagine if you were even a retailer and suddenly your customer relationship and your customer data is actually being sucked into one of the Google, Amazon, or Facebook intelligent agents. You’ve now suddenly been disintermediated And you are now a supplier or a spoke in their massive digital ecosystem. And you’ve just completely lost your competitive advantage in the world. And this is, this is the disadvantage of the new neural networks coming out. So most of the brands that my company is working with at the moment, and the reason why I’m actually up here in the UK is that the big brands that understand this are actually working with us to white label our technology and bring out intelligent agents that serve their customers, but enable them to flip the model, share the data that they know about the customers back with them so the consumer is in control of that data. And more importantly, the enterprise brand is not just an intermediate, but in fact can become a master of their own digital universe through these intelligent agents. So it’s replacing ultimately the website, which is actually pretty dumb and static. And in fact, rather than having 15 apps to, you know, have recipe planning, have food delivery, and do my shopping, the intelligent agent can manage all of that. And it’s centered on me and, you know, what my dietary preferences are, as an example. The type of coffee that I like, because a good intelligent agent will always remember you and what you like, and that’s exactly what we do.

Speaker H: Right, but isn’t there going to be an incentive on behalf of the large technology companies though, to try to dupe your intelligent agent, to try to play with the neural networks and to divert it into an area that’s more in the interest of a large technology company and its search capabilities than necessarily in your interest?

Speaker I: Well, not necessarily, because intelligent agents are quite clever, because what they have is they have neural networks and they also have what we call ethics and rules. And so, they know the parameters that they work within. In fact, an intelligent agent is less likely to missell a product or a service than a human. So, you also have authentication and security there, so you know who you’re talking to. There’s a lot written also about chatbots. There’s an awful lot that’s been put out there, and they’re a dangerous wee thing, in my view. Because chatbots are coming out and they could be working for, you know, the Russian mafia, and you might think you’re talking to a human being, but it’s not a human being on the other end, and you could in fact end up being duped on, you know, that you’re, you know, you’re talking to somebody that is not working on your behalf or is working for, for, you know, nefarious purposes. And again, this is where it’s important that you— that, you know, from our perspective, you set up trust networks with brands and you know who the intelligent agent’s working for and where your data is going and that it’s under your control. And we did an Ipsos independent survey in the US, and 86% of American consumers said that they would not use an intelligent agent if they didn’t know where their data was going and it wasn’t under their control. And that’s what’s very important, that you have transparency, trust, and control for the consumer with these intelligent agents. And then of course, from a brand perspective, that they actually also have some, some trust and control there as well. And this is where there’s a big arms race going on between the big tech giants and —like the Googles and the Facebooks—who think that they would like to suck up the world’s data and have Google deciding what’s in your fridge or what the temperature of your home should be versus consumers or people being in control of their data. And that’s a big debate going on with the British government and with the European Union at the moment.

Speaker H: That’s going to be an issue, though, isn’t it? Given the amount of times that your data could be used by an intelligent agent, obviously that will be very, very difficult for you to to check or to continuously monitor that? How can you be sure about that movement of your data?

Speaker I: Well, very good question. So first of all, before you interact with an intelligent agent, check out who they are working for. Where is that data going? And if you don’t know, then you probably shouldn’t work with that intelligent agent. Our point of view from MyWave is that we issue our consent receipt. So what happens is that when our intelligent agents is now working. Let’s say for example, I’m working with a brand and now my intelligent agent wants to deliver a new television to my home. My intelligent agent will arrange for the purchase of my television, will arrange for it to be delivered, but then once that TV has been delivered, if I don’t wish to have any more relationship with that delivery company, that consent receipt is revoked and that data is actually bought back. So in fact, the data is no longer with that agent. So we have the ability to track where your data goes and we use permission and consent receipts. So these are some of the new things starting to come out in the industry so that people actually can see from a transparency perspective where their data is. And we put, where we’re very different from my company’s perspective is that we actually have an intelligent agent that learns about you and all that data goes into your personal cloud.

Speaker B: Hmm, my own personal cloud. That sounds attractive, as long as it’s a small cloud, not threatening to bring climate change. But neural networks remind me of the ghastly scenes in the film Ghost in the Shell, the new one with Scarlett Johansson. In a dark, dripping basement, human beings sit in a circle, comatose, their brains linked by a mesh of cables to provide processing power to the rogue human who is trying to upgrade himself by preying on others. Of course, that’s not what is meant by an actual neural network. It’s just a set of linked servers. But the monster-in-the-machine myth still has a powerful hold on my imagination. So an IA, intelligent assistant, can use an AI, artificial intelligence, to help us take back control of our data, to help us buy things and services. But creative people also need to sell things, to get them from A to B, and to build a following for their music, art, films, writing or other artistic creations. In this respect, the online world is forced to mirror the real world with national boundaries, different languages and currencies, and a range of restrictive practices that get in the way. Consumers have the option of downloading them for free using BitTorrent sites like Pirate Bay, which still exists in spite of a series of legal actions against the people behind it, and totally legal video-on-demand services like Netflix offer blockbuster content worldwide. In order to safeguard the creative and cultural diversity that exists at the moment, decision-makers in the UK and European Union are trying to devise rules that will be fair to producers as well as consumers. Angela Mills-Wade, Executive Director of the European Publishers’ Council, has been explaining to Pete Warren that this is actually a difficult and delicate task.

Speaker G: At the very beginning of that whole debate, when they put this forward as a proposition, copyright was often blamed as the reason why content was not available in different countries. When I say content, I mean a particular film wasn’t released on the same date that it was perhaps released in London, or wasn’t broadcast on prime-time television in one country, whereas it was in France, or whatever. Gradually, the European Commission and other people realized that actually it was much more complicated than that, that there were other reasons perhaps including the online retail chains, not least Amazon and Apple, had their own way of dealing with things country by country. So the kind of hated market segmentation that the European Commission talked about was actually a function of practical things like people’s credit cards are linked to where they live, and often the online retailers don’t like moving out of that kind of comfort of knowing that the person who’s ordering is paying with a card that’s linked to their address and all that sort of thing. So there were kind of practical considerations of the retailers. And it is true that certain types of content, especially audiovisual content, is sold. The rights are sold on a market basis, territory by territory.

Speaker H: But people do, if you’ve got some subscription to Netflix or something like that, people do want to watch that, you know. So I quite often go to Leipzig, and I actually I do have a subscription to Netflix, but I wouldn’t watch it. But I’m just saying that, you know, you may wish to watch Netflix if you are in Leipzig, for example.

Speaker G: And I think that— and I do think that’s perfectly reasonable, and they did in the end deal with that separately. So there is a regulation on what’s called portability, which comes into force, which means that if you are a lawful subscriber to something, which doesn’t necessarily mean you’re paying for it. So for example, BBC iPlayer, you know, as a license fee payer, assuming you are one, you ought to be able to access the iPlayer whilst you’re outside the country as well. And I think the regulation stipulates various conditions for it, but in a nutshell, you will soon be able to do that with your subscribed services. But what the producers are worried about is not that so much, but that if, if it’s kind of open sesame for anybody to buy anything on Amazon at the cheapest price, and whilst there are, you know, other differences due to market conditions, that then it becomes a bit of a bazaar and it’s not really, not really fair anymore. And going back to the original point about undercuts the financing, the pre-financing of large-scale productions. But the portability thing is a sort of halfway house, and everybody thinks it’s perfectly reasonable that if you’ve paid for something, you shouldn’t really be blocked from enjoying it just because you’re in another country. You know, copyright incentivizes investment in production because it means you can then say what what is done with your stuff. It gives you control. The interventions in the past have, have been to incentivize local production. It’s to make sure that the quotas for independent production versus state broadcast finance production, European language production versus imports from America or Asia. It’s all been about maintaining and supporting cultural diversity. And even those in the book trade, interestingly, the big, big companies in the book trade will tell you that licensing on a global basis is fine. But it does tend to, especially in the English language, lead to a kind of eye for bestseller. There’s kind of bestseller syndrome that, you know, do you take on smaller authors if you have no idea if they’re going to be any good. So I think there is a fear in the audiovisual sector that if you move towards this one market, you then lose the, the kind of special nature of cultural diversity. Because if you’re a broadcaster and you’re forced through regulation to change your business model, you’re going to go for the kind of blockbuster syndrome again.

Speaker B: Angela Mills-Wade. And as ever, technology is marching ahead of the lawmakers with a new system just launched on the market for tracking copyright breaches using, you guessed it, blockchain. Berlin-based Copytrack aims to build a global registry to help its clients get properly paid for the use of their intellectual property using its own currency system of tokens.

Speaker J: Sandro Maeda of Copytrack told me how it works. At the moment, sometimes photographers use several agencies to distribute their images, and sometimes these agencies are not actually very accurate in how they report these sales to their suppliers, so the photographers. So at the end of the month, they simply get a report which states like it was sold 4 times, but They don’t tell where it was sold to and how it will be used. And this is actually a problem because right now we find a lot of results and the authors of the image are then not quite sure if this result, so this finding which we have online, is actually unlawful use or whether the person or the business actually licensed the image because they have no idea who the image was licensed to.. And in the future, by integrating also a marketplace into this blockchain which can report exactly when an image was sold to who, it makes it transparent for the author itself. He can then see straight away, ah, my image has been sold to this and this webpage or this and this user, so he’s allowed to use it and doesn’t have to worry about that anymore. These are actually the two main focuses on how the blockchain can help us with our business in the future.

Speaker B: What’s your business model?

Speaker J: How does Copytrack offer a return on investment? So right now what we are doing with the ICO is actually we are selling the tokens, so CPY tokens, which is our cryptocurrency which will be used in the whole ecosystem of our blockchain as a payment. So in the future whether it’s the registry or whether it’s marketplaces, this coin can be used to buy licenses for images or to register images. So in the future, there will be a higher demand on these coins, and this will obviously influence also the price of the coin itself in, for example, dollars or euros. Right now, what investors can do right now, they can buy this token for lower than what we expect it will be in the future once all these businesses are set up which will be using this coin as payment within the system.

Speaker B: Technically speaking, how do the Copytrack crawlers identify images that might be breaching copyright?

Speaker J: Right now what we do is, this technology is divided into two parts. First, the crawlers, which will go through the internet to several web pages and will simply look at the images and report them back to us. And then the second part of the technology is the image matching engine, which will look at the images which our customers provided, so the original images, and will match them with the images which we found online. And once these matches, once these comparisons show a match, they will report it will be reported to our customer and will show up in his account as a hit. And then he can decide whether this is actually illegal or not legal. So our technology simply actually looks at the images and compares them to the images of our customer. And once we found a match, we show it to them and they decide illegal or not illegal.

Speaker B: Of course, being able to create and enjoy music, movies, TV, and radio shows in different languages is one of the joys of being human. Yet here too, the machines are marching in, with translation apps in every traveller’s smartphone and language-learning possibilities that involve only one human—you yourself. Password reporter Charlotte Roundtree has been testing out a new system from Rosetta Stone, a long-established language school. What’s new is that it claims to mimic the way we all learn to talk as babies, in a natural method. Panos Koraniotis explains how it works.

Speaker A: You know, when we talk about technology and learning, especially in the corporate space, so learning itself has evolved. What’s important is that yes, technology allows a business to make learning more efficient and effective because it’s more accessible, you can access from anywhere, anytime. It makes it more scalable so you can touch more people. In a far easier way. The logistics of coordinating everything online is far easier than having all these on-site classes. But then more specifically, you can use the technology to make it engaging. Now, that said, and we’re true believers of this, technology on its own is not enough for effective learning. And what do I mean by that? There’s two other things that are required specifically to language. One is pedagogy. So, the way in which you teach has to be academically sound. You know, there’s a lot of companies out there that have simply digitized language learning, and some companies think that, “Hey, I’m going to take on this digital tool and push it out to my population, and language learning is going to flourish.” That simply isn’t the case. You have to continuously strive to improve language learning with the use of technology, not simply digitize it. You know, 30+ years ago, our founder had learned German, and then he embarked on learning Russian, and he failed miserably. And it was because of the methods he used, but more specifically, it drove him to ask the question, you know, why are we so able to master our first language so well? And the conclusion he came to was that it stems down to the way in which you’re taught. So, if you compare the two, your first mother tongue you learn in a full immersion-based environment, and you do so over, you know, whatever, 5 to 7 years. But what is that environment? It’s an environment where there’s no translation, there’s no grammar rules, there’s no conjugating of verbs. It’s literally trial and error and interacting with the surroundings around you in a very immersive, interactive way. Based on that, he developed this principle that if we can use technology to recreate the way you learned your first language as a child, then not only are you learning how to think in the language without translating, you’re learning how to understand the language. You will figure it out. And more importantly, because you’ve worked it out for yourself and no one’s told you what it is, especially not a translation device, then it’s far more likely that you’ll remember it. There’s a lot of psychological studies that show that if you solve a problem yourself or discover something yourself, you’re 90% more likely to actually remember it. So, it’s all in the instructional design and the pedagogy of what we created to— like I always tell people, we’re not a digital language company. Some people make that mistake of thinking we’re like comparison to Babbel or Duolingo or any of these. You know, we are pushing the boundary of language innovation. We’re trying to improve the way that the world learns languages to make it more efficient and effective.

Speaker G: What’s the main draw for English-speaking UK residents to learn different languages?

Speaker A: So why do most people seek out things like Rosetta Stone? There’s, I guess, various answers to that question. If you’re looking at it from a personal perspective, it’s more than likely a significant other who’s from a different international background than yourself and you’re trying to connect with them or their families. But from a more commercial perspective, from a business perspective, it’s always been providing yourself with further opportunity, either in career or just in general life experience. If you look at a different demographic of population in terms of younger population, it’s normally backed by general international curiosity, which is fantastic to see in young people because What that does is it just provides them with a far richer perspective of the world. So the various reasons come from all angles. What’s for sure is that the byproduct of what’s achieved by becoming multilingual is that you definitely widen your own horizon, whether that’s in opportunity or in simply how you look at the world and how you envisage it.

Speaker G: So you don’t think realistically there could be an upsurge of just translation software that’s coming out recently as opposed to language learning?

Speaker A: Yeah, and it’s a fair comment. I think there’s always two ways of looking at something, especially when it comes to technology and how it will evolve the way in which we approach something. I think to answer the question, it’s important to always think about what’s the purpose of speaking a language and what’s the intention behind it, and that’s to communicate.. And in order to truly communicate with someone, there needs to be this human touch, this human element of it. So, yes, you can very easily, like you did 20 years ago, walk into a hotel lobby and repeat what is written to you on a page, which is a translation page, very much so like one of these devices would. Or even simply having a translator with you. However, the ability to truly build a relationship and in some way, in a very human way, connect with someone has to come from your own efforts to speak that person’s language. And for that reason, I think that certain people will find that they will use the translation vacation technology route, maybe to order a cerveza on the Costa del Sol perhaps, but people that are more interested in connecting with others while travelling and understanding cultures and doing business and generally kind of looking outward as an individual internationally, you need to be learning the language and reap the benefits that that brings. So some of the research that you can find in terms of what language learning does to smaller children and, you know, improves decision-making capability. I’ve already mentioned it gives you more perspective in the world, that there’s not just one way of doing something. It gives you stronger relationship-building capabilities, but also I think there’s even evidence that shows that it helps in multitasking. You know, there’s loads of benefits that can come from learning what is a code, another code, another way of communicating.

Speaker G: In theory, distance learning is never going to match up to having a teacher in front of you. You only need to Just think about pronunciation. Books can’t teach you to pronounce letters and sounds that you don’t have in your own language. And you can’t hear yourself answering a tape. So how does Catalyst overcome? With simple voice recognition tech, the kind most of us have on our phones. Converting your words into sound waves so you can see visually where you’re going wrong and where the highs and lows of where your words should be is much easier than hearing the differences for yourself. And when that isn’t enough? Catalyst links you in real time with language speakers to improve your skills in real conversations, giving you that human side of communication as well. It’s a clear showcase of technology that’s aware and looking to teach properly, and it’s as meticulously planned and organized as a classroom, from its lessons on grammar and structure to the vocabulary it filters in slowly. The success of the software is obvious and very positive, but what does it mean for teachers and classes in today’s schools?

Speaker A: Already, you’ll be glad to hear, Charlotte, that already we do a lot of work with schools and teachers. And what’s important in the teaching environment is that this was never designed to replace teachers. It’s really important I emphasize that. Sometimes we come across teachers that might feel a little threatened by us, and that was never the case. A teacher will always have a role within a school in terms of guiding that student through what is a minefield of information. Where we excel in the education space is 1 hour a week is not enough time for a kid to learn a language. There are so many distractions and so much going on that just engaging that skill set for that little time in your working week is not enough. So where we’re very predominantly used is providing additional exposure to the language outside of the classroom. So once they leave the classroom, they get extra exposure maybe every 1 or 2 or 3 days That way when they come back, not only have they been excited by their own progress and capability which they see in themselves in the programme, they’re also more engaged in the classroom because they’re not just coming in and out of it once a week.

Speaker B: So machines can teach us natural language, and of course they can also learn it and share the knowledge with other machines. 4 years ago, a project launched that aimed to teach a home entertainment system all about human emotions. The EmoSpark Cube exceeded all its crowdfunding targets and is now linked to NASA satellites. It is programmed through hard encoding according to the so-called Laws of Robotics first devised by the science fiction writer Isaac Asimov in 1942. Intelligent machines must respect organic life and prioritise humans. Asimov Laws, of course, were devised in a fictional context And policymakers, as well as technology companies, are still grappling with regulations for the real world to protect humans against the machines. For the rest of us, the 200th anniversary year of Frankenstein’s monster should prompt us to make a resolution to get our digital data under control. Whether it’s an intelligent assistant, a HAT, a personal cloud, a virtual private network, or another solution, The important thing is to know who’s got our data, whether it’s accurate and up-to-date, and how they’re using it. Oh, and to have some fun with Frankenstein.

Speaker E: You probably recognize my name. I’m the fifth generation of scientists in the Frankenstein family. Sometimes with a name like Frankenstein, it’s difficult to be taken seriously, thanks to that book. But here at Life at Kids Lab, I’ve decided to embrace my legacy. We’re devoted to exploration. Exploring what makes us human, whether it’s our biology or our culture. And we’re looking for people, especially young people, who are curious. People who like to do things, make things, ask questions. And we’ve partnered with 50 science museums around the nation. We know that you We know you have busy lives, so we won’t take much of your time. We’ll just ask that you check in every 3 days or so for 30 days of science and maybe a kind of frank-adventure. Sign up at frankenstein.life to start the experience.

Speaker B: Password is a Future Intelligence production for Resonance FM with Peter Warren, Charlotte Rowntree, Blue Buffery, and me, Jane Wyatt. You can find out more about technology and society at our website futureintelligence.co.uk. We’ll be back next month with the Resonance FM fundraising auction where you can bid for some really exciting experiences.

Speaker A: Thanks for listening and goodbye. This programme 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.