This week: Mona Lisa’s fake smile, more Digital Humans, and missing platform competition. Sandra Peter (Sydney Business Insights) and Kai Riemer (Digital Disruption Research Group) meet once a week to put their own spin on news that is impacting the future of business in The Future, This Week.
The stories this week
00:45 – Mona Lisa ‘brought to life’ with deep fake AI
07:43 – Join us at Vivid Ideas 2019 for a discussion around Digital Humans
09:07 – Platform competition: why privacy is an antitrust issue
Other stories we bring up
Salvador Dalí deepfake at a Florida museum
Joe Rohan’s synthetic voice used to recreate podcast
Listen to Joe Rogan’s fake podcast here
Google assistant can translate in your own voice
Bella Hadid and Lil Miquela in latest video for Calvin Klein’s “I Speak My Truth in #MyCalvins” campaign
Our previous discussion of CGI ‘influencers’ like Lil Miquela
Our previous discussion of hype and digital humans
Our previous discussion digital Mike
Our previous discussion of Xinhua News Agency’s AI anchors and BBC newsreader who speaks languages he can’t
NYT opinion piece by Facebook co-founder Chris Hughes
Tik Tok is the emerging Chinese competitor in the social media space
Apple’s recent App Store antitrust court decision
Chinese company forced to divest its share in dating app Grindr
Our previous discussions of #BreakUpBigTech
Kai’s comments on should Facebook be broken up
China’s arguably the world’s most dynamic start-up scene
Our previous discussion of the TikTok phenomenon with Barney Tan
Our 100th episode looking at LilNasX, TikTok and the #OldTownRoad phenomenon
Come see us at Vivid
Join us live at our Vivid Ideas 2019 event on June 7 at the MCA!
You can subscribe to this podcast on iTunes, Spotify, Soundcloud, Stitcher, Libsyn, YouTube or wherever you get your podcasts. You can follow us online on Flipboard, Twitter, or sbi.sydney.edu.au.
Our theme music was composed and played by Linsey Pollak.
Send us your news ideas to firstname.lastname@example.org.
This transcript is the product of an artificial intelligence - human collaboration. Any mistakes are the human's fault. (Just saying. Accurately yours, AI)
Disclaimer We'd like to advise that the following program may contain real news, occasional philosophy and ideas that may offend some listeners.
Intro This is The Future, This Week. On Sydney Business Insights. I'm Sandra Peter, and I'm Kai Riemer. Every week we get together and look at the news of the week. We discuss technology, the future of business, the weird and the wonderful, and things that change the world. Okay let's start. Let's start!
Kai Today on The Future, This Week: Mona Lisa's fake smile, more Digital Humans and missing platform competition.
Sandra I'm Sandra Peter, I'm the Director of Sydney Business Insights.
Kai I'm Kai Riemer, professor at the Business School and leader of the Digital Disruption Research Group. So Sandra, what happened in the future this week?
Sandra Well lots of dead people seem to be coming back to life. One of the stories on BBC this week was that "Mona Lisa has been 'brought to life' with deepfake AI". So we're talking of course about the very famous Mona Lisa painting, Leonardo da Vinci's Mona Lisa, which has now been animated using artificial intelligence by a group of researchers coming out of Samsung's AI research lab in Moscow.
Kai The BBC reports that of course it's not just the Mona Lisa, the researchers also brought back to life Salvador Dali, Albert Einstein, Fyodor Dostoyevsky and...
Sandra Marilyn Monroe.
Kai Exactly. So this is the latest instalment in the rise of the Digital Humans. We have of course talked about this on a number of occasions before, just about a year ago we talked about Lil Miquela for the first time and she's making headlines this week as well. We've talked about AI news anchors in November 2018. And we of course had a special on Virtual Mike, which we aired last year off the back of our Vivid event in 2018. Now, bringing back to life these personalities, and a picture for this matter, raises a few questions.
Sandra The easiest one to answer is really: how have they done this? Because while the technology that researchers have used to animate Mona Lisa really relied on training an algorithm to understand facial features and the way the face moves and the shapes that it makes. And they trained that on 7000 images of celebrities from YouTube, and then they use that knowledge to animate one still picture of Mona Lisa. Also over the last two weeks we've seen quite a few stories about the Salvador Dali deep fake. And this is different to the way Mona Lisa was created in that there were many, many actual recordings of Salvador Dali, who died in 1989 that were used by the Dali Museum. This is in Saint Petersburg Florida, one of the museums with the largest collection of surrealist works. They've debuted a human-sized Salvador Dali on a huge screen, where you can actually interact with a replica of Salvador Dali.
Kai So Salvador Dali actually speaks, fully-animated, a virtual replica of the man himself, talks about his life, his work, poses for selfies with the museum patrons, and then texts to those images. So, quite a lively installation that supposedly brings back the real Dali.
Audio I do not believe in my death. Do you?
Sandra As you can hear, they've also reproduced Salvador Dali's voice. Which makes this almost indistinguishable from real recordings of Salvador Dali. Which brings us to the second question, not only how can you do this technically, but who gets a say in what Salvador Dali says after his death?
Kai So as we now gain the digital technology to bring back to virtual life people who are deceased, such as celebrities, painters, movie stars, politicians, the ethical question that underlies all of this is how do you recreate the personality? How do you decide what this person would have said, when we can now involve these digital humans, replicas of actual people, in conversations, in installations, talking about their life, and go places in conversations where these people might not have gone in recordings?
Sandra So you could use this for educational purposes. It would be wonderful to study history alongside Napoleon or from Winston Churchill. But on the other hand, there's also the potential to misuse this technology...
Kai And to recreate history, basically. We all know that we relate well to other people, to faces. And when we recreate these digital humans, we're creating a very convincing medium through which the author, who creates these installations and controls these virtual humans, can talk to audiences, and large audiences potentially on the Internet not just in museums. To then have people, basically, say whatever the author wants.
Sandra Which brings us straight to another story that came out last week in which AI recreated the voice of podcast host Joe Rogan, and actually did make him say anything they wanted.
Audio I've decided to sponsor a hockey team made up entirely of chimps. I'm tired of people telling me that chimps are not capable of kicking human ass in sports. Chimps are just superior athletes, and these chimps have been working out hard. They're throwing kettle-balls, battle ropes, everything. I've got them on a strict diet of bone broth and elk meat. These chimps will rip you...
Kai Oh shit. I think we better stop this here.
Sandra Technology is now so advanced that with only a few seconds of someone's voice, and admittedly there's hours and hours of Joe Rogan out there, he's the host of one of the most popular podcasts ever.
Kai Usually episodes that are three and a half hours long so anyone complaining about the length of our podcast, just saying.
Sandra The point here is that you could make anyone say anything you wanted, and indeed my students in my class have created the version of The Future, This Week with me asking questions that I've never asked before using free software available on the Internet.
Kai And clearly we need to upgrade our tech here so Megan could have created an episode last week when we both were unavailable.
Megan Well, that would be easier.
Kai But the point here is that this technology can be used in all different ways. There was also a story of Google and Google Assistant, which can now translate what someone says, and use their voice to actually speak the translation. The same technology being used as in the Joe Rogan example.
Sandra And one more story on this topic again this week from People magazine which reported on the Calvin Klein ad that faced really intense backlash. This was a Calvin Klein ad featuring human model Bella Hadid...
Kai And Lil Miquela who we've featured a number of times on the podcast before. The fully synthetically rendered Instagram influencer who has millions of followers taking fashion and life advice from, in quotes 'her'.
Sandra And, who in this latest Calvin Klein ad, titled "I speak my truth in #MyCalvin Kleins", is seen sharing a long lingering kiss with Bella Hadid.
Kai Which is technically significant because here we for the first time have the fully digital human, so to speak, interact with the real human in a video, and so the boundaries between what is fake and what is real are quickly disappearing in audio recordings, as in Joe Rogan, in bringing back deceased celebrities, or in having fully synthetic humans now appear in videos alongside real humans. And while we have seen this previously of course in movies and CGI, we are increasingly in a world where we can do this in real-time, off consumer technology.
Sandra And this is exactly what we will be doing at Vivid Ideas next week. We urge you to join us, there are a couple of tickets left. If you want to see this technology live and an in-depth conversation of some of the issues that we've just touched upon in this podcast, join us at the Vivid Sydney on June 7, 5:30 pm at the MCA, Museum of Contemporary Art. Join us for an in-depth conversation, a showcase of these cutting-edge technologies., and a panel of experts from around the world.
Kai It will feature Michela Ledwidge, Mike Seymour (the real Mike, not the Virtual Mike. Hao Li, from the University of Southern California in LA, and Sandra and myself. So come along to a in-depth discussion of the ethical, societal, business, economic implications of living with Digital Humans.
Sandra And enough of that, I think we should move to our second story of today which was a fascinating one.
Kai This is a topic that we've kept our eye on for the past few weeks and we've always thought 'oh, we can't talk about big tech and Facebook again because we've just done a couple of stories around this topic'. But it turns out that there is an angle which, frustratingly for Sandra and myself, has been missed a number of times now in reporting. We thought 'we have to talk about this'. This one comes off the back of yet another opinion piece in The New York Times talking about Facebook. Mind you, there's been a very widely discussed one a couple of weeks ago by Chris Hughes, one of the Facebook co-founders who called for breaking up Facebook, and made a wide ranging argument about why Facebook and the lack of competition that it faces has led us down this dark rabbit hole that is big tech these days. But the latest one makes the argument that privacy itself is an antitrust issue.
Sandra So in the article we're starting with today Dina Srinivasan, who is an antitrust scholar, argues that there is an antitrust case to be made against Facebook on privacy grounds. Privacy, unlike things like price gouging or price discrimination or exclusive dealing, privacy is not one of those things that you would immediately think of or recognise as an antitrust violation.
Kai So this harks back to a discussion we've had previously on the podcast about how the thinking around monopolies and antitrust law in the US is slowly shifting, and we've featured Lina Khan, a post-graduate student who has written a influential thesis around the topic, recovering an older, almost forgotten interpretation of antitrust law and the role of monopolies. Which argues that it's not just the price effect for consumers which in Facebook's case, because the service is free, would never actually come to bear. But that other issues can also feature in antitrust thinking. And so the argument goes to privacy.
Sandra So the author makes the case that Facebook has usurped our privacy with the help of its market dominance. The argument goes that whilst the price of Facebook has always stayed the same, it's always been free to join and it's always been free to use, the cost of using Facebook to us, and that would be calculated in the amount of data we have to give up to Facebook in order to use the free service, would be an order of magnitude over what it used to be. So today when you sign up to Facebook you have to allow it to track your activity across, the article reports, more than eight million websites and other mobile applications on the Internet. And the thing is that you cannot opt out of this tracking. You can opt out of what Facebook can do with that data, but not of the actual tracking.
Kai And it goes so far that even if you leave Facebook and shut down your account, Facebook will keep tracking you across the Internet and collecting data about you.
Sandra It's actually even worse because the moment you do leave Facebook, you no longer have access to the settings that would tell Facebook in what ways it can use your data.
Kai And so Dina makes the argument that while price is not the factor here, that this invasion of privacy, and the way in which Facebook works with that data that it collects about us, amounts to a degradation in service, or quality of the service, which then goes straight to the antitrust argument.
Sandra And also harks back to many of the points that Hughes, one of the founders of Facebook, that he raised in his article where he highlights over and over again the fact that there is this lack of competition that has enabled Facebook to become so complacent and really, disregard the wishes of its users...
Kai Or squash competition among startups, which is another angle that has often been brought up when it comes to Facebook and big tech more generally.
Sandra And here is where we want to pause.
Kai Because, surprisingly to the both of us, the main point that the Hughes article, a number of other articles that have appeared since, and the latest one, seem to be missing is that the privacy invasions that we find on Facebook are not the result of a lack of competition, but the result of quite intense competition.
Sandra So let me say this again: we're arguing that actually it is not the competition that a company like Facebook would face from other social media networks, from Instagram or from WhatsApp. It's actually the competition based on their business model that we should be looking at. And in that respect, Facebook faces really, really intense competition, and the breaches in privacy that we're seeing are a direct result of the competition that it faces.
Kai Once we acknowledge what the product is that Facebook sells for money, and everyone knows this, it's advertising, it becomes pretty clear that there is intense competition with Google, as the main competitor, and that the way in which the product that is being built, the targeted advertising that is being sold to corporations, depends directly on how well Facebook can actually learn everything about their users, hence exploit their privacy, it becomes clear that competition is the actual driver of privacy invasions.
Sandra And I feel this is almost a big enough insight to stop the podcast right here, but we want to go a little bit further. So what we're saying here is that the issue that we need to be looking at is not the lack of competition, but the very intense competition, hyper-competition at the back end of these platforms. And the answer to that is not breaking them up, but rather regulating them.
Kai Which interestingly, might also be missing the point because there's yet another layer of competition that seems to have been overlooked.
Sandra And this is again an argument for there being a lot of competition. None of these articles actually go into discussing the other big tech competitors. The conversation seems to be restricted to social media companies and platform services that are coming out of Silicon Valley, completely missing the fact that there are Internet giants of that size, some of which are direct competitors, coming out of China.
Audio MUSIC AND LYRICS of Old Town Road.
Sandra We're talking of course about companies like TikTok. We've recently done a podcast on this. TikTok is a social media network with as many users as Facebook.
Kai And is currently the most popular social app in the US. And it is of course coming out of China, owned by Chinese company ByteDance.
Sandra And of course more on this in our recent 100th episode of The Future, This Week where we discuss the story of L'il Nas X.
Kai And have recently also done the yeehaw challenge, we'll put this silly video in the shownotes.
Sandra But coming back to the intense competition, all of the discussions about breaking up Facebook, breaking up Amazon, completely miss the fact that there will be competition from companies that are rapidly growing out of China. And just to give you some numbers there...
Kai So big stats coming up.
Sandra Just over the last five years, China has produced more than 34 unicorns, that is private companies valued at more than one billion dollars. And this is fuelled by a number of things, chief amongst which, there are 700 million users and climbing using the Internet in China. There is a flood of private venture capital investment, and a huge, huge state-backed funding and state-supported incubators assisting this ecosystem in China.
Kai So companies such as ByteDance with TikTok, but also increasingly WeChat and others are expanding their business beyond China into places like Africa, Australia, the western world, the US, Europe more broadly.
Sandra And think not just social media companies, but also companies like Didi which have taken on Uber successfully in many jurisdictions, or companies like Huawei which are actually at the centre of very big trade debates at the moment.
Kai So for Facebook this means that, rather than looking at it in a narrow way, arguing that it is a lack of front-end competition in the market for social media that is at the heart of the many problems that we see. We should recognise that not only is the competition at the back-end the driver for privacy abuse, Facebook is also increasingly confronted with new competition from Chinese competitors that actually operate with a very different business model.
Sandra Which also makes them extremely well suited to compete with companies like Facebook and Google in this space, because whilst the company like Facebook would have over 98 percent of its revenue coming from its advertising business, a company like Baidu or Tencent would have only a very small amount of their revenue coming from advertising. The rest would be from financial services that it offers from entertainment and so on, so a much more diversified business model which is a lot more robust in terms of its exposure to competition for advertising revenue.
Kai So what we're saying is there's a whole new angle to this whole big tech conversation emerging, the implications of which we're only now beginning to understand. And we want to finish on two observations here, now first of all TikTok, because of the different business model does not face the same issues as Facebook regarding privacy invasion.
Sandra Or the same level of scrutiny.
Kai On the other hand US regulators might be made uneasy by the fact that this is a Chinese owned company that now is part of the lives of millions of US citizens, or indeed citizens in other countries such as Australia.
Sandra And there has been a recent case with a Chinese company acquiring Grindr, which is currently being reversed in the US courts.
Kai So in 2016, the Chinese company Beijing Kunlun Tech bought 60 percent of US dating platform Grindr, completing a buyout early last year. But the government has now decided that this transaction is actually a threat to US national security and hence Kunlun is expected to shortly sell Grindr at an auction. The point being here that the Grindr app for example allows location tracking of people, finding matching people on the basis of their preferences. And the argument being made that should Chinese owned companies have ownership of such data about US citizens, that might raise national security issues.
Sandra So it remains to be seen what is the case with companies such as TikTok, which are fully-owned Chinese companies. And it also remains to be seen what are the economic consequences of breaking up big tech in the West, where there is significant competition from places like China.
Kai So platform competition will be a topic that we will return to in future episodes, and not only because of the Chinese angle, but also because platform competition turns out to be a much more interesting and multifaceted topic than is often portrayed in straightforward antitrust opinion pieces, for example.
Sandra But I think that's all we have time for today. Again, come see us live at Vivid on June 7th at the Museum of Contemporary Art. Enjoy an evening of ideas and lights.
Kai See you soon...
Sandra On The Future...
Kai Next week.
Sandra This week?
Kai Yes, but next week.
Sandra On The Future, This Week. Next week. Thanks for listening.
Kai Thanks for listening.
Outro This was The Future, This Week, made possible by the Sydney Business Insights Team and members of the Digital Disruption Research Group. And every week, right here with us, our sound editor Megan Wedge who makes us sound good and keeps us honest. Our theme music was composed and played live on a set of garden hoses by Linsey Pollak. You can subscribe to this podcast on iTunes, Stitcher, Spotify, YouTube, Soundcloud or wherever you get your podcasts. You can follow us online on Flipboard, Twitter, or sbi.sydney.edu.au. If you have any news that you want us to discuss, please send them to email@example.com.