This week: winter is coming, Uber knows you’re tipsy, and take the call. 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:
Other stories we bring up:
Future bites / short stories:
Our theme music was composed and played by Linsey Pollak.
Send us your news ideas to firstname.lastname@example.org.
Disclaimer: We would 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: winter is coming, Uber knows you're tipsy and take the call.
Sandra: I'm Sandra Peter, Director of Sydney Business Insights.
Kai: I'm Kai Riemer, professor at the Business School and leader of the Digital Disruption Research Group.
Sandra: And we're back!
Kai: You're back. I was here.
Sandra: Yeah, I'm back from South America and getting we're back into the news so Kai what happened in The Future, This Week?
Kai: Our first story is from VentureBeat and it's titled "The AI winter is well on its way". So, winter is coming apparently.
Sandra: Winter is coming and even though Game of Thrones Season 8 is only coming out next year, the AI winter seems to be coming sooner.
Kai: The term 'AI winter' was coined when the first installment of AI in the 1970s hit a brick wall and AI became rather unfashionable after the hype of the 70s what is now referred to as good old fashion AI. Back in the day when we tried to teach the computer the world and tried to implement logic reasoning in machines, things didn't quite work out the way they were supposed to. And AI fell into a long and dark period.
Funding dried up and only some really committed researchers inside institutes and universities kept the field alive. And there's now fears that the same might happen after the current hype comes to an end.
Sandra: So, a few people have, and this article comes on the back of that, people have been starting to see the cracks in the hype around AI. And as you mentioned this idea of an AI winter of not only reduced funding but of reduced interest as well might be coming for this current wave of deep learning that has been at the forefront of this current wave of AI. And we've talked about it on the show before, many people have been talking about the inevitable coming of the singularity, people over the last few years have been touting the current wave of technologies as world changing that they will fundamentally revolutionise a number of fields and the article gives as an example the self-driving car technologies, that companies like Tesla have been announcing that for the past few years have been said to be almost here, to be revolutionising, life changing and yet have somehow failed to come. Not only that but increasingly we're seeing what the limits of that technology is. And quite recently again we've talked about this on the show, we'll include the episode in the show notes, the crashes like we've seen with the Uber car that has unfortunately killed a person.
Kai: So, Tesla actually been selling fully self-driving capability in its auto autopilot technology even though we haven't even invented it yet. So, the article takes that as one example of the over boarding hype, the hyperbole around AI. We've heard stories both good and bad of the robots coming and solving all our problems and climate change and disease prevention and all kinds of different fields and vice versa the narrative around the robots coming for us. The late Stephen Hawking and Elon Musk calling it the greatest threat to humanity and so the article makes the point that a lot of this hype has really been unhelpful because it has raised expectations to the point where it's likely never going to be fulfilled especially not in the short term and the applications of AI might be rather mundane making AI that looks promising in the lab, work in the field can be much much harder. The results can be much less convincing and so people who are deeply involved in the field working on these technologies have come forward and said these things are really much harder and much slower than many people have predicted and so there's a fear now that a lot of the investment going into the sector might never actually yield any returns and that the public might lose interest and that we might actually be headed for a period where AI and the field as such becomes discredited.
Sandra: And as you have mentioned before this wouldn't be the first AI winter. There was one at the end of the 70s. There was another long AI winter at the end of the 80s and through most of the 90s until we came up with new ways of doing AI and also such winters are not constrained to this particular set of technologies, we've seen similar things for instance with railways in the mid-1800s in the UK for instance where we had the railway mania. This was said to be a technology that will change everything.
Dozens of companies were funded, very few of them actually made it. Very few of them made good on the technology which was still in its infancy. Large pools of investment actually dried up when technology failed to deliver as quickly as it was expected and as spectacularly as it was expected which is the same thing we're seeing now with AI. But critically we want to ask one question here: why do we have this big hype in the first place. Why do we feel that we must have these stories of fundamental revolution transformation of singularities coming of the end of everything? It's be said to be bigger than electricity and all of these things why do we have the need to have these stories?
Kai: Yes absolutely. This is not a story about AI, this is a story about technology more generally and the nature of the hype cycle. And to me it's actually the story about how sometimes we see a very pronounced hype cycle, not necessarily just technologies moving along the hype cycle, there's overblown expectations, then things cool down and they become more productive but rather that when technologies becomes so overhyped that things crash down to the point where as much as the hype went up the disillusionment comes down and the winter notion really stands for a cooling down to the point where no one really wants to be seen to be involved.
So the example is the dot com boom of the early 2000s when every company, every investor everything had to be E or dot com and once things crashed in terms of expectations not coming through but also financial losses, share market crash, a complete dry up of IPOs, it actually for a while became toxic to have the dot com in the name and many companies were quick to abandon it and to reposition, to pivot as we say today and many many startups just disappeared. So, the question is are we headed for this in A I. Not a question we want to engage with because predictions are not really what we engage in here on the podcast we'll leave that to other people. But we want to ask the question why do we actually end up with this hype in the first place.
Sandra: Because of a number of reasons actually. One of them simply has to do with the fact that we like a really good story. Such stories are actually important in the beginning of these hype cycles to get the amount of funding that you would need to actually do research of the scale that is required for artificial intelligence. But then there is also the media. Killer robots make for a very good story and the Terminator makes for a very good story.
Kai: Yeah we like a good scare right?
Sandra: We also like a good news story. We will be able to cure cancer, we will be able to get to Mars. We will be able to have self-driving cars or flying cars that will drive us around.
Kai: And there's also what I called the religious belief in technology - we are so accustomed to technology solving our problems that we actually want to believe the stories about technology that progress will be infinite. Progress will be exponential. That we will replicate human intelligence in machines so we're very ready as a society to believe the stories about technology and when this hype is in full swing it is very hard for people to get heard when they have a counter argument. I've been called a technophobe on Twitter by my own personal trolls arguing against the hype of AI and we've been talking on the podcast for almost two years now trying to put in perspective what's happening, highlighting the real advantages, the real dangers.
Sandra: And how difficult progress is in fields like that.
Sandra: How many people actually have to work on this and for how long. When we look back into history and we look at other industrial revolutions we sometimes forget the length of time and the huge periods of time that pass from the invention of the actual technology. Things like the steam engine to the rise of railway networks, of companies that would span continents and so on. That takes place at a different time scale than what we're expecting to happen at this point.
Kai: The problem that I have with this hype is that we're spending and wasting a lot of money, a lot of funding that goes in to ideas where it is pretty clear to experts in the field that they will never actually come to fruition because expectations are overblown, promises cannot be kept because the technology is just not up for it. But a lot of money is being wasted that could go to much more pressing and much more real problems when this hype is in full swing and any kind of bullshit idea receives funding.
Sandra: Yes absolutely. And maybe one more thing before we go to our next story which is to say that even though an AI winter might be coming and we might be getting as far as we can with the current set of technologies, let's not forget that special purpose machine learning, special purpose AI is actually making inroads in banking, in law, in medical diagnostics. So the AI revolution as put forward by people like Elon Musk might not be coming but we might still be seeing large scale transformation and a leap forward in both technology and the way our society is organised. Let's not forget that whilst we had the first industrial revolution around the steam engine, the technology itself it took the next wave, the second industrial revolution to actually start seeing besides the advances in traditional fields like engineering to see advances in the rise of large scale corporations, in new socio technical systems, in medicine and in law. So it took more than a hundred years to see new type of structures emerging.
Kai: And indeed something similar has happened in the wake of the dot com boom. While the topic was discredited and university departments who had rebranded as disciplines of electronic commerce quickly went back to their former designations, slowly but surely the Internet has transformed the way in which we do business - Amazon has become by far the largest retailer or indeed one of the largest corporations in the world. So even though the topic itself might drop off from the media and the hype might be over, that is not to say that those technologies will not unleash their transformational potential over time and that might actually be quite a healthy state of affairs which lets people go back to doing actual research rather than becoming celebrities that are handed around in the media. So while we're not engaging in any prediction there are now signs that this hype might be coming to an end.
Sandra: And our next story is actually an example of the hype and of people finding new uses for a technology much like a hammer looking for a nail.
Kai: Oh a very nicely done segue Dr Peter. The next story concerns Uber and their use of AI to predict whether you will dear listener when you are calling your Uber on a Saturday night might have had one or two drinks too many.
Sandra: Our second story comes from the BBC and its titled: "Uber applies for patent to spot drunk passengers". The patent application that Uber has just filed actually tries to spot on unusual behaviour and it weighs in a couple of things. Things as how many typos did you have while trying to type in the address of where you're going to.
Kai: How long does it take you to actually master calling that Uber?
Sandra: How precise are you when you press the In-App bton to actually call the car?
Kai: Are you walking more slowly than usual?
Sandra: Are you standing around on the kerb? So someone who might be mixing up their words or swaying from side to side or taking 15 minutes to manage to place the order in the app is probably intoxicated if this happens to be a Friday night or a Saturday night.
Kai: So let's unpack this. Now first of all what is a striking reminder again is the extent to which companies that provide us with these apps can actually monitor what we're doing. They have precise information of where we're walking through GPS data. They can monitor the speed at which we're typing so they basically have access to many things that we do on our phones as we're doing them. They can read messages before we actually send them because they can observe anything we do in the app. And the idea is here that monitoring the user behaviour in that way can tell us something about the state of that user. Or so it says in the patent application.
Sandra: So let's look at what it would mean if companies such as Uber or indeed a taxi company were to implement this. So first let's remember the Plan B campaign from Transport for NSW. This involved having a plan to get home safely after a big night out. It was highlighting that driving yourself is not an option and taking a taxi or these days an Uber was one of the suggested possibilities for making it home safely.
Audio: If you're drinking don't drive. RBT means you need a plan B.
Kai: So for our international listeners RBT stands for random breath tests. So the whole point of calling an Uber might be that you have been drinking and that you are not driving.
Sandra: Now in this case the app could actually be used to identify inebriated riders and the Uber driver choose not to pick them up.
Kai: Which would defeat the purpose of calling the Uber in the first place.
Sandra: So in this instance, maybe not such a good thing. You would rather have someone pay for cleaning up a car than have them die horribly in an accident. On the other hand let's think about the way in which the algorithm identifies uncharacteristic user activity, swaying side to side, not standing up straight, taking a long time to order the car. These could be characteristics of someone who is inebriated but also someone who has a certain disability, maybe suffers from Parkinson's disease.
Kai: Which comes down to the way in which such a technology might be implemented. So if indeed the algorithm learns, so takes as the baseline behaviour the way in which I normally use the app and then deduces any deviations from that, sure that might work, but if a user's behaviour is simply compared to the average user's behavior we might discriminate against anyone who actually uses the app differently. People with disability.
Sandra: But even in your own case right it might happen that you are actually in need of something, something has happened, you are having a seizure.
Kai: Yeah I may be having a stroke, I might really need some help and the Uber app is the only thing I could muster.
Sandra: So there are domains in which maybe we don't necessarily need AI for that?
Kai: And there's actually been other articles, this Uber story has been all over the news, that have asked some serious questions about how this might be implemented and who might have access to this information? Pointing at the rather sketchy history of Uber and its drivers when it comes to for example of sexual harassment. So The Guardian asked is it such a good idea to give Uber drivers access to a feature that might let them identify female users who might be intoxicated? Potentially creating a feature that would allow people to prey on people who are defenseless.
Sandra: So all in all may be not the best way to think about how Uber could benefit from machine learning, artificial intelligence.
Kai: To me, because it's a patent application, it might be one of those cases where a technology is looking for a problem to solve. And one final point, another article pointed out that drivers in a forum were discussing that this might not be anything nefarious or dangerous, it might just be Uber looking for a way to make more money because if you're intoxicated we might charge you a higher fee or we might actually charge you a s urge price and you might not notice because of your current state of mind. So it might just be one of those features that allows Uber to inflate its prices.
Sandra: As you mentioned, for now just the pattern, but we'll keep an eye out on how this is developing. And to our third story for today comes from the Atlantic.
Kai: Titled "Why no one answers their phone anymore" and Sandra and I've been discussing this a bit. Sandra's rarely taking her phone calls, happy to let them go to voicemail while my habit is still to take almost every call unless the caller I.D. tells me that I really shouldn't.
Sandra: So that's true. You've even stopped this podcast to take calls.
Sandra: So the article points to a number of reasons why the telephone culture seems to be disappearing. And mentions, as probably the most important aspect which is quite a structural one, that we really have a lot more options in how we communicate with each other now, even on my phone I've got a number of apps from WhatsApp to Viber to Hang outs, We Chat, different ways of engaging with other people. So voice is now complemented by text messaging with all its variations, with emoji and gifs and regular photos and videos and links and everything else that comes with it. You've also got ways to talk to a number of people at the same time, you can post things on Twitter, on your Slack, on <phone ringing> this is one of those calls I am ignoring because I see it's Kai calling.
Kai: I thought I'd get a word in. So what I find interesting about this article is how beautifully it outlines how it used to be an absolute norm and convention that when the one telephone in the house rang, we would actually take the call. Someone would have to dash to the phone, even though it was bolted to the wall or sitting on a desk somewhere, and take the call because you never knew, when first of all, who might be on the phone, in the days before caller I.D., but it was also polite to take the call and it would have been a no-no not to answer it, and times have changed significantly, and communication is changing and it is still in my view an expression of the norms in our culture, which are changing as well on the back of these technologies.
Sandra: And let's not forget, not only on the back of the technology but also on the practices that have grown around those technologies, right now probably 90 percent of the calls that you're getting is someone trying to sell you a new Foxtel plan or a new electricity plan or asking if you're happy with your banking services or your car insurance or whether you want to buy property in Nigeria.
Kai: Yeah, yeah very true. A lot of calls these days are in fact spam. I find peculiarly, there's one profession where everything seems to be happening via phone, so journalists are usually calling me. So that's one reason why I sometimes just pick up the phone. For some reason people in this profession never email, they always call, which I find quite interesting really. So interestingly different parts of society have different norms, but my observation is that our communication has become much more geared towards efficiency, information exchange and we have taken out the human element, and if we look at some of the recent inventions or announcements...
Sandra: ...You're talking about things like Google Duplex?
Kai: Oh absolutely. Gmail compose, where we now have AI trying to predict what we want to say in an e-mail, helping us be more efficient in creating text blocks for us to compose messages. Again, taking out personal touch, just creating messages for us, our electronic assistants talking to each other to make appointments and create calendar entries. So a lot of the inventions seem to be geared towards reducing the human element making communication more efficient and more about the information exchange.
Sandra: But at the same time you can argue that we're trying to find new ways of putting the human element or the emotional element back into it. If you think about emojis or the reaction gifs that you can send to people with cats giving you a hug or high fives or.
Kai: Arguably that's not a very efficient way to communicate. Okay. Well I hadn't thought of that. I still think were struggling with communication just for the sake of it especially in a work context. But it is interesting to see that there's many many diverse new forms of communication and obviously gif emojis are one way of communicating.
Sandra: And another thing I want to add here is that this seems also to be somehow culture specific, because whilst we're seeing a decrease in voice communication in countries like Australia or in places like the U.S. we've actually seen over the last few years a huge rise in voice communication in places like China where apps such as We Chat actually enable voice messaging and are used more for voice messaging than they are used for text messaging. So phones are in this case used basically as walkie talkies for sending short audio clips instead of t ext, so entire communications are basically exchanges of little voice clips and people don't use headphones they just hold up the phone to their mouth and speak in it and listen to it on the phone speaker. And this is done quite publicly and in large groups.
Kai: And interestingly Apple has now announced a new feature for its Apple Watch, something that I've actually long been waiting for. On the back of observing that this is happening in WeChat, the Apple Watch will get a so-called walkie talkie feature where you can send short voice messages to someone else who has an Apple Watch to engage in those short back and forth voice messages using the Apple Watch as a walkie talkie effectively.
Sandra: So, I think it will be quite interesting to see whether this catches on because whilst in China, this idea of having a loud and lively culture is actually a sign of confidence and a sign of being important. If you're doing this publicly and if you're being loud and listening to your messages, in places like Taiwan where Mandarin is also used where you could make the same argument it's hard to type in Mandarin, people are not really using the same voice facilities that We Chat that enables there. Similarly, whilst in India you have a number of different dialects and maybe it's easier to say things rather than to type them in, such features are not used on apps like WeChat or WhatsApp. Whilst in Argentina, and you'd argue Spanish is about as easy to write in as English, the feature is widely used on WhatsApp. So interesting to keep an eye on how this plays out.
Kai: And yet again a reminder that communication media are a reflection of our social norms and that technology itself does not dictate but rather shape the way in which we communicate. So, let's finish up with some future bytes. So, what is something you learnt this week?
Sandra: And now for something completely different. I came across an article in The Guardian that looked at the split between sugar and solar in Australia, up in Queensland. And this article talked about solar farms spreading across agricultural regions of Queensland and this tension now between people who previously grew Sugar Cane who are now giving up a lot of their productive agricultural land and leasing it out to companies that are now building solar farms on them. So interestingly renewable energy has made forays into places like Queensland, still relatively small ones, a very small part of Queensland energy comes from solar, but they haven't popped up in areas where you would normally expect them, where there is land that is not used for anything else, but instead they have shown up in areas where you already have an industry present, the sugar industry, in this case in Queensland, because some farmers have been struggling with growing these but also these lands are quite close to infrastructure that already exists making it a lot easier for the renewable energy industry to set up solar plants in these areas.
Kai: So that's quite interesting, that actually ties in with another article that I saw just recently outlining the problems that Australia is increasingly having from the amount of solar that is being brought online and we had the special during your holidays where we reported that China is also having problems that the grids on a good sunny day especially during the day as people are not at home, their solar is producing a lot of energy. But no one is actually using much of it, that the grid finds it hard to cope and we have to invest in grid connected battery to actually store that energy so that it doesn't overload the grid. So, it's quite interesting to see how solar has become an economic proposition, not one to subsidize to actually speed up technology adoption but a genuine way to generate new income sources in farming. So, it's going to be interesting to see how this plays out from here as we are reaching solar saturation, while we still see lots of big projects coming online.
Sandra: So what was your story for this week?
Kai: Mine is an unlikely candidate concerning cauliflower. So why would we talk about cauliflower of all veggies on The Future, This Week? Turns out cauliflower has become big business, and this comes on the back of a whole flurry of innovation in the food tech sector. We've discussed clean meat startups that try to come up with lab grown food. Turns out there are also low-tech solutions that are making their way into the marketplace. So, cauliflower turns out is a great ingredient when it comes to producing gluten free and veggie-based replacements for all kinds of different food groups. The article reports about a woman, a marketing executive with two kids with celiac disease who for a long time was after recipes to basically cook for her children, and she found a way to create cauliflower pizza dough, and she's now heading up a multi-million-dollar brand Caulipower which markets pizzas deep frozen and pizza dough across nine thousand stores in the US. And the article goes on to outline how cauliflower is really a great ingredient to create alternatives to rice, noodles, all kinds of different ways to replace food groups that normally contain things that are not great for people with celiac disease but also tie in with a health trend where gluten free has become a trend not just among sufferers of this disease.
Sandra: I actually like cauliflower. So, I think this is a good place to end because that's all we have time for this week.
Kai: What about broccoli?
Sandra: I like broccoli too.
Kai: Yeah. So here on The Future, This Week...
Kai: We actually like broccoli, we like carrots, we like cauliflower, we like all veggies but this is all we have time for today. The food space is an interesting one and we will keep an eye on this. Thank you for listening this week.
Sandra: 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 from a set of garden hoses by Linsey Pollak. You can subscribe to this podcast on iTunes, Stitcher, Spotify, 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.