The wine we drink and machines that can think
It's that time again where we round up some of the most pressing and perplexing science news stories from the past month. We analyse the first images from the James Webb telescope, discuss whether the protocol on rescue from car wrecks is based on any scientific evidence, and ask how data might be used to the detriment of period tracking app users in the wake of Roe vs Wade. Then, we'll find out why great white sharks are running scared near South Africa, question whether chatbots are on the verge of becoming sentient, and compare how the taste of wine is affected by the glass it is carried in...
In this episode
First Pictures from the James Webb
Chris Smith
The James Webb space telescope launched on Christmas Day last year; it reached its final position in late January about 1 million miles away. It’s at what's called a LaGrange point, which is a region of neutral gravity that makes it easy for the telescope to park itself.
The telescope mirror that collects light is about 25 square metres in area, which is at least 6 times bigger than the one carried by the Hubble telescope and means we can see far further than we've ever seen before.
And the really exciting thing about the James Webb is that it's built primarily to see infrared light, or heat.
That might sound counter-intuitive: why would we want to see heat rather than look at visible light that we normally can see with our eyes?
The answer is that this will enable the James Webb to peer through the fog of gas and dust that spans the Universe and often obscures other colours of light, including visible light, which means that - previously - some distant parts of the Universe have been off limits to us. But because infrared can thread its way through that material fairly easily, we now have the equivalent of powerful night vision goggles that can show us what's really out there.
There's also another advantage to relying on the infrared: as the Universe has aged, it's expanded. This has had the effect of stretching the rays of light coming to us across space. This is called cosmological red-shifting. And it's most pronounced for some of the oldest - and hence earliest - galaxies that formed in the young Universe.
The light from those has been red-shifted to such an extent that it now sits in the infrared, so the James Webb is also a window back in time to what the Universe looked like perhaps just a couple of hundred million years after the Big Bang. This means we can use it to test our theories of how the Universe has evolved and grown.
And perhaps most exciting of all, because different chemicals soak up and emit unique colours of light, by watching as distant planets float past in front of their parent stars, we'll be able to see - from the colours of the light they block out - what the atmosphere is made of on those distant worlds. We'll be able to find would-be Earth's - and discover how common our world is - in the grand scheme of things.
It’s certainly a very exciting time, and the word from NASA is, predictably, watch this space!
03:36 - Could more people be saved from car wrecks?
Could more people be saved from car wrecks?
Tim Nutbeam, NHS
Meanwhile, back down here on Earth you probably assume that our emergency services, those tasked with getting us out of harm’s way, were always working with the full weight of scientific medical evidence behind them; if so, this story might come as something of a shock. This is because, in the case of rescue from car wrecks, research has outlined that the current approach, designed to minimise the movement of patients to guard against exacerbating spinal injuries, is counter productive. And it’s not marginal. In over 99% of cases, spinal injuries are not the main problem for the patient, while extending the amount of time they spend trapped in the car could render other injuries, where every second that treatment is delayed counts, potentially fatal. Speaking with James Tytko, Tim Nutbeam, an Emergency Medicine Consultant for the NHS, estimates that hundreds of people in this country, and many more worldwide, might have been saved if a scientifically driven approach had sooner been applied to car wreck rescue efforts…
Tim- The principles of current fire and rescue techniques are based around what we call absolute movement minimisation. So, when you suffer an injury to your spine, there is a concern that further movements might exacerbate that injury and make the primary injury worse, and we call that secondary spinal injury. That has been the paradigm, the focus of extrication techniques and how they've developed, particularly over the last 40 or 50 years.
James- What are the flaws with this approach?
Tim- It's a, a single consideration in a complex environment, and we need to have an understanding of what injuries patients have and make sure that we've weighed all those into balance to work out what is best for that patient and that particular extrication. We set out to fill in some of the gaps in the relevant scientific evidence, and then to re-evaluate - is what we are doing correct? And if it's not, what are the next steps?
James- What was the result of that re-evaluation? What did you find?
Tim- So we went back and looked at all the history, all the farm rescue manuals, and tried to work out where this paradigm of movement minimization had evolved from, and if it had a historical evidence base. We could find no evidence of that. There was never any underlying scientific evidence in terms of frequency of injuries or type of injuries, or a specific reason why that, that paradigm might have been chosen over, say, a head injury or a chest injury or an abdominal injury. The second part of the project was looking at the different injuries which these patients suffer and we worked with the trauma audit research network and is the biggest trauma database in Europe. What we did is we looked at patients who were trapped and patients who weren't trapped and then looked at what injuries they suffered and then reported those injuries and also used modelling techniques to see if being trapped led to more deaths than it should.
Tim- What we demonstrated was the rate of spinal injury, or spinal cord injury, was very low, less than 0.7%. All the other patients - so this is looking at majorly injured patients - had other significant injuries. Normally it's the head or the chest or the abdomen with spinal injuries right down at the bottom of this list of injury frequency. We also found that being trapped was associated with more deaths than it should be. So once we took into account people's age, their comorbidities, that's the other things that they've got wrong with them, and the injuries that they suffered, we found that being trapped was an independent predictor of mortality. So there's something about being trapped that kills you.
James- Is it fair to ask why this has taken so long for people to look at the way we deal with these accidents and to re-assess how we should be dealing with them?
Tim- Yeah, I think there's a number of different challenges to that. One is around, I guess, patient ownership. Who is responsible for casualties or patients when they are actually trapped? Are they actually a patient or are they a casualty which is the responsibility of the fire and rescue services? And I think these areas of shared ownership lead to potential difficulties in delivering research in that area.
James- I definitely appreciate what you're saying, but when you've outlined the situation as clearly as you have, how does it stand up to people out there who, obviously have my deep sympathies, maybe they've lost someone to an accident who might have stood a far better chance had their injuries been treated sooner, giving an updated approach, which moves away from the movement minimization way of dealing with things.
Tim- I think that's a difficult question to, to answer. I think the important thing is, is that we've identified that what we are doing could be better and we are working with all the right stakeholders to make sure that that information rapidly and efficiently translated into practice. It's a shame that we've not looked at this sooner, but now that we have, we seem to have captured momentum and the buy-in and support of all these stakeholders to ensure that this translates rapidly into not only guidance, but into the practice that we deliver on a day to day basis, moving forward. We've already seen some changes or results and, and that's really fantastic to see.
James- Yeah, that's good to hear. When it comes to this transition, are there any major sticking points in the way that you envisage or could, could things move quite quickly from here?
Tim- I think it's important not to underestimate how difficult it is to overcome years of institutionalized practice and the fire rescue services have been taught for many years, that even a very small movement can lead to catastrophic consequences. I think it's important that this work has, it's not delivered by doctors and it's not delivered by fire rescue personnel. It's not delivered by a certain group of practitioners, but it's been genuinely developed and delivered by a multidisciplinary group of stakeholders. I think, in the past, we've perhaps gone wrong by doctors telling people how to do things. And I think the principles of regularly revisiting this guidance and ensuring that what we're doing is patient centred and patient injury centred that we can continue to refine and get better at this area of practice and all those other areas of shared multidisciplinary practice.
09:42 - Are period tracking apps a privacy concern?
Are period tracking apps a privacy concern?
Andrea Ford, University of Edinburgh & Laura Lazaro Cabrera, Privacy International
At the end of June, the legislation which was interpreted as conferring the right to have an abortion in the US - Roe vs Wade - was overturned, leaving the decision about the legality of the procedure to individual states. Since then, half of US states have since made moves to ban or restrict abortions, which technically means any woman seeking the procedure could be prosecuted. In the wake of Roe vs Wade, viral messages spread quickly online, telling individuals to delete any period tracking apps they were using. But why are these apps a concern? Julia Ravey spoke to Andrea Ford, a medical and cultural anthropologist at the university of Edinburgh…
Andrea- There have been a lot of concerns raised since the news in the US. The concern is that the data in your app would be potentially used as incriminating evidence, if you're in a situation where you're being persecuted for having an abortion. So that could include data about missed periods, abortions, miscarriages, irregular periods potentially. Phones in general, and lots of apps, track location data which could be incriminating, if you visited an abortion clinic.
Julia- All these concerns valid. Do you think there is a chance that data could be used in this way?
Andrea- Yes. I think there are definitely valid concerns. Lots of app companies have stepped up to be reassuring, which is nice that they're recognizing the concerns. There are, however, a lot of loopholes or different ways of interpreting privacy policies in legal jargon. Sometimes if you're protected in general, you're not protected if you're under criminal prosecution, because that's a special set of rules. While it's nice that app companies are often recognizing this. I wouldn't necessarily say that it's entirely reassuring because there's lots of ways to get around regulations and promises if there's the will to prosecute.
Julia- And these laws being overturned, it comes at a time where we now have these devices essentially logging everything we do. It's very different to the pre Roe versus Wade world. Do individuals need to be thinking about their data beyond just period tracking apps?
Andrea- Yeah I'm glad you asked about that beause I think, in a lot of ways - not just related to abortion - but in privacy in general, period tracking apps have been this red flag. People have become concerned about them because the type of information you put in is quite intimate. But, they're just a Canary in the coal mine in a sense there's so much data gathered and surveillance of different sorts that is baked into the digital ecosystem that we live in. As I mentioned earlier, the location data that's shared and stored by any number of other apps might actually be more incriminating in the court situation.
Julia- From person to person, periods vary dramatically. People can have a regular period but stress and incidences like that can alter periods. Do you think that using the timing of someone's period could be used to incriminate someone because of this huge variation?
Andrea- Again, I think that comes down to the will to prosecute. There have been numerous historical examples where people have said they've had a miscarriage and have been prosecuted for an abortion, and there's just a grey area between the two. And if you look historically, the way similar experiences have been interpreted has really varied. So because there's this ambiguity inherent in the thing, that's why I say it comes down to the will to prosecute. If some judicial body, jury, et cetera, is bent on interpreting it a certain way, that will matter much more than the actual data.
Julia- What would be your advice to someone tracking their period using an app?
Andrea- At this point, I think it depends where you live. I live in Europe and I kind of have a handful on the go for research purposes. But if I lived in a lot of areas of the US, I would not. There are apps that store your data locally, instead of in servers. Those are more secure. Although your actual phone could be held as evidence in a prosecution situation. So, if you're concerned, I would say don't use an app. The kind of takeaway I would love to leave people with is that considering your personal privacy is important and it's part of a bigger set of problems and having conversations about those bigger questions is really, really necessary at this moment.
Julia- In order to learn more about how our data can be used in criminal proceedings, particularly in the overturning of Roe versus Wade. I spoke to Laura Lazara Cabrera, a legal officer from Privacy International who, as an organization, look at the exploitation of data by government and organizations. She told me about the types of evidence that might be used to get a conviction in criminal proceedings.
Laura- If you have indicated in your engagement with a period tracking app that you are pregnant or that you are looking or have undergone an abortion, then that information is likely to be of interest to law enforcement authorities and, by extension, prosecution to show that you have indeed undergone such a procedure. Of course, it's not just period tracking apps. You could have shared that sort of information with a range of different actors. It's all about tracing back your steps and seeing who you told and what sort of engagement those private companies are likely to have with law enforcement. In other words, what is the likelihood of them passing on that information?
Julia- And how easy would that sharing back be? So say you've got an app on your phone and you agree to their privacy measures and they've come out and said, your data is protected in some way. In the case of criminal proceedings, can the user privacy guidelines be overridden?
Laura- Privacy policies will usually have an exception, or at least they should when this is the case, that law enforcement will be able to access some of that data or at least that they will be providing some data to law enforcement upon law enforcement making a legitimate request. But sometimes the issue is that this is not prominently flagged in the privacy policy and so people may overlook that. I think that everything that's happening in the US should prompt users to reconsider their terms of engagement with private companies and look at the privacy policy where there is one.
Julia- Have there been any examples of this in proceedings? So, prosecution using data from an app to get a conviction.
Laura- So we know that browsing histories have been used in the past in the context of criminal proceedings and data processed by period tracking apps is just something that could fall in the same category in terms of the types of evidence that that could be. Then again, whether or not that is likely to be conclusive evidence of someone having undergone an abortion is a very different question altogether. But the real worry is that combined with other evidence or in light of other information, this type of data could be incriminating and put a person in a very difficult position to defend themselves.
Julia- Some of these period tracking apps are based in the EU, would a person in the US using one of these apps, therefore come under EU data protection.
Laura- Absolutely. So the great thing about the general data protection regulation, or GDPR as many of us know it, is that it's got a really broad scope of application, which includes people based in the territory of European union, but also companies which are based here. The good thing about GDPR is that it provides quite a strong standard of protection for people in terms of the ways in which their data is processed. Of course in the US, we don't have such legislation. Each state has different types of legislation that may be applicable to data protection but, in the end, there is no single overarching federal piece of legislation that looks into data protection.
Julia- So in the US, are you more vulnerable in these situations? What is the difference between the EU privacy laws and the US in terms of protecting personal data?
Laura- The US simply doesn't have that sort of legislation in general. There is some legislation that looks into health data, but then that legislation only applies to certain healthcare providers. So that means that many of these organizations will be following through the cracks in terms of data protection standards in the US. Whereas, if we're looking at the EU, then it doesn't matter whether or not you have been licensed to provide a certain service everybody's held to the same standard.
Julia- What would be your advice to someone who is maybe based in one of these states now where reproductive health is being restricted in terms of tracking their menstruation using an app, or just in general with their data being on their phone?
Laura- I would say to them to treat everything that's happened as a result of Roe v Wade as a wake up call and try and adapt to the current information environment. Take this as an opportunity to go back to those apps that you've been using for a while, with whom you've been sharing, very intimate, personal data. Review the privacy policies, make sure that you are happy with those privacy policies. If you're not, try and find out what you can do in terms of maybe limiting your engagement with the app, if you feel uncomfortable continuing to use it or exercising the data protection rights that may be included in the app.
18:49 - What is hunting down the great white shark?
What is hunting down the great white shark?
Alison Towner, University of Rhodes
Now, speaking of the underwater realm, Gansbaai in South Africa was renowned as the Great White capital of the world; at least until 5 years ago, when the animals disappeared, almost overnight, leaving researchers and locals puzzled. Since 2017 Alison Towner, from Rhodes University on the Eastern Cape, has been piecing together the different parts of the jigsaw puzzle to work out what’s happened…
Alison- It wasn't as if we saw a decline, you know, they just vanished. And initially it happened for about six to weeks and the sharks started to reappear again. What had happened in that time period was kilo whales had arrived and been here for 15 years. I'd never seen killer whales really in this area. Certainly I'd never seen the type of killer whale that's known to hunt sharks. So this is actually quite a rare morphotype of killer whale that specializes in selectively targeting sharks and their livers and, lo' and behold, here were these two very distinctive killer whales. So they were here in the study site and they proceeded to predate on various great white sharks in the area. We know this because the carcasses of the dead white sharks washed out and, in the wake of that, everything changed. So post 2017, when this happened, you know, the numbers like literally drastically declined and then the white sharks would sort of return again in much lower numbers. So, with the killer whales, and it was just a game of cat of mouse since then, until here we are five years later, the, the sciences finally published and we can show the world the evidence and the data. And it's, it's quite startling.
Harry- Is this the first time that we've witnessed this behaviour, this predation of the killer whale eating the, the great white or white sharks.
Alison- This behaviour has been documented once before off of San Francisco, California, at the Farallon Islands and scientists there, it's a 1999 paper so it's quite a while ago, described observational account of some offshore killer whales killing a great white shark. Biting it at the surface, towing it, pushing it through the water to asphyxiate it, and then sort of rolling on its back and, and extracting its liver. There was a written report, it wasn't direct evidence and it certainly wasn't any carcasses washing out for examination afterwards. So, lo and behold, what happened here was the carcasses were very coastal because these types of killer whales are generally more offshore. So the carcasses washed down, that was the first time in the world. We were able to sort of post mortem them. And I actually led all the necropsies, we've had eight so far, but yeah, having been part of that now and seen so many of them washed out is quite spectacular, but quite sinister at the same time, the killer whales literally rip them open at their fins and then just extract the liver and discard the carcass.
Alison- It's incredible.
Harry- Do we know why they're just taking the liver?
Alison- Well, the liver is a huge organ, very lipid rich, so very fatty nutrient high profile nutrient. It's just, it's right there for the taking in that if they flip the shark over and put it into a trance like state and they tear it, then the liver literally is so accessible, right? We've estimated that male adult killer whales would need to eat about one white shark liver a day to sustain itself. So <laugh> the calculations have been done. I mean, it's not gonna exactly fill them up for very long, but they certainly specialize in extracting this one organ.
Harry- Is it that these sharks then know that this is going on. And so that's why they're leaving the area or there's certain cues that they're taking before they run off, away from Cape town.
Alison- So if a white shark's taken off guard by a killer whale - if all of a sudden it turns around and there's multiple killer whales - I mean, we know there's at least two that are doing it. The sheer trauma of that have being chased of other of the same species in your area within, you know, range of you detecting it being predated on. It's probably enough to drive you out the area for one. But then it can be the rotten carcass. So, once it's discarded and it's laid on the sea bed, let's say it hasn't washed out, we do believe there's more that hasn't washed out. The scent of death, if you will, could keep the sharks away, but that would be momentarily, not long term. So all the questions are there, but somehow these white sharks have this evolutionary built in mechanism to know to avoid these <laugh> these areas where it's high killer whale risk. I'm sure the science will definitely unravel some more about that with regards to vocalizations of killer whales or exactly what mechanisms the sharks use.
Harry- I'm making the assumption that at the moment, you're kind of at the start of this bringing together of data. So with that in mind, you know, everybody will be aware that food webs and ecosystems are very fragile, they're very complicated. If you take away something like the great white shark from an environment where it is an apex predator, what happens to the rest of the sea floor? Do we know what happens to that environment?
Alison- At this point? Of course it's speculation, it's hypotheses, but I mean, removing top predators is never good for ecosystems full stop. And now the coastal bay where these white sharks sort of reign king, if you will, is home to so many other critically endangered species that are actually reliant on the white sharks being there as top predators. So, for example, we have the Greater Dyer island system here in Gaansbai, which has critically endangered African penguins on it and their numbers are right down. They can't sustain any more pressure. And now, because just adjacent to their colony is actually a colony of breeding cape fur seals. Now the seals don't have their natural predators around. So they're actually directly competing with the penguins for food, but also predating on them. So again, it's just that little perturbation at the top tier of the ecosystem in our Marine environment here, that's now out of balance and it has cascading effects right The way through the whole food chain. Yeah, If we leave it too long and you know, this isn't rebalanced, then we could have serious consequences in our hands. But what I will say is that we have a couple of great white sharks that have finally come back to this region. There's about four of them out there today that the boat's seeing, but it's been almost a year since they were here.
25:23 - Wine taste altered depending on bottle colour
Wine taste altered depending on bottle colour
Will Lowe, Cambridge Distillery
Storing white wine in showy flint glass bottles could cause the wine to deteriorate, a new study published in PNAS suggests. Italian researchers tested over 1,000 bottles to see if the colour of the glass had any effect on the contents, simulating what happens before your drink hits the supermarkets. Here to explain the findings, Will Lowe the founder and master distiller of Cambridge Distillery. If you think 1,000 bottles of wine sounds like a lot to sample, Will has tasted over 10,000 in a year because he’s also a master of wine. The lucky Harry Lewis went to see him at work…
Will - Fundamentally, the idea of a bottle is to get the liquid to you, the consumer, in as close to the form and condition that the winemaker intended as physically possible. In many situations, that means that it should taste, feel, look, present exactly as it did when it left the winery, and in other situations, it can be quite different in the case of wines that are intended for maturation, for example. And then we have the secondary role, which is that these days, very importantly, the bottle is part of the self-marketing of that liquid. It's there to help sell it. So regardless of whether a wine may or may not improve within the bottle, if nobody buys it, no one would ever even know. And I think particularly in the case of rosé wines, it's become increasingly important that people are buying with their eyes. Without people being able to see that, that would be a very hard sell indeed.
Harry - So it actually does have an effect depending on what you bottle your wine in can influence the taste? How does it do that?
Will - This has been well known for many years; the first person to write about what's now known as light strike was written back in, I think, 1983. So what we're looking at here is two potential mechanisms of breaking down of flavour. Clearly, you have heat properties that can be exacerbated through light, but the real key that we're looking at here, and the one that's coming under some scrutiny, is the influence of UV light in particular on wine and the flavours therein. The flavours of wine, made up by both volatile, aromatic compounds (the things that you can smell), but also its physical characteristics, the alcohol and the acidity etc that are latent within the bottle. As a bottle of wine changes - and all wines under all conditions will always change as this is a dynamic biochemical environment that we're looking at - you can see an emergence of favourable or unfavourable aspects, as well as a reduction in both favourable and unfavourable. What we're looking at in this study specifically seems to be predominantly the deterioration of the favourable characteristics and in particular, those which are offering varietal typicality. So that flavour that makes a Sauvignon Blanc so easily identifiable, for example, appears to deteriorate under certain conditions with greater exposure to UV light.
Harry - Any excuse to have a midday drink, I've bought a couple of bottles along - two sauvignon blancs that I just got from the local supermarket. So again, I should probably really stress that this isn't a forte of mine. I do go for the lightest rosé though, but I just think it's because it's in trend.
Will - It's very Instagram friendly!
Harry - It is Instagram friendly. I wonder if perhaps we could taste one and then you might be able to describe what would happen if we were experiencing light strike, how that flavour would change.
Will - Sure, let me fetch a couple of glasses here then... Now the theory has it that light strike would be most common in clearer glass. So in terms of filtering out that UV spectrum or close-to-UV spectrum that we want to avoid, green glasses slightly better, amber Glass is better than that, darker brown is even better yet. The harder it is to see through, the harder it is for the UV light to get through. You've very kindly brought two in here. One is a slightly light green colour, which we'll try first. And then the other one in as close to clear as you're likely to be able to find, I would imagine. Now, Sauvignon Blanc is famously most often produces a single varietal, meaning there's just one grape type in there, as opposed to, for example, in Bordeaux, where you'll find a combination of different grapes. For that reason, most often labeled varietally - so it'll say "Sauvignon Blanc" - and then, fairly common new world themes is that you'll have, as you've got here, a name of a random animal; we call them "critter wines". They're the cute naming after an animal is usually what sells them. What makes Sauvignon Blanc so identifiable to most people is it has this very distinctive aroma. Not everyone can smell that, there are genetic predispositions that will mean that you can't, but for those of us that can, it sits somewhere between gooseberry, fresh cut grass and passion fruit, I think would probably be the closest and easiest way to get to it. Gooseberry is most often used as the descriptor. That comes from a certain set of volatile aromatic compounds, which we are seeing through this paper are susceptible to this feature of light strike. I've not even tasted that, I've just nosed it, and this is part of building up a familiarity; you do get to a point where a quick nose of the glass is enough to help you with the varietal identification. If we pick up the second glass, to be honest, this is working better than I'd expected it to do. Certainly, you can still sense on the nose alone the aromatic profile that would lead you to Sauvignon Blanc, but it is muted, isn't it?
Harry - The clearer glass bottle is muted? It's not quite as punchy?
Will - Yes. If I were to use a fairly crass metaphor, I would say that the darker glass is turned up to eight out of 10 volume and the clear glass is maybe a three or a four. It's a marked difference. Now, all things are not equal here. These are from two different countries. It isn't a fair single variable assessment that we're making here, but that is demonstrably the case, that this does smell far more restrained. Let's be honest, that could also be winemaker intention. Some consumers will want a vastly more focused and aggressive style, and some people will want something that little bit more elegant and demure perhaps. But what we saw in this paper is that they ironed out that crinkle of winemaker intention by taking the same wines and treating them differently. So they were able to really isolate that as a single variable.
Harry - Let me jump to a big conclusion then that we definitely shouldn't do. If there was a take home message for other listeners, would it be stick to dark bottled wines?
Will - No, I think it's really important that we have this onward march of science to help explain things that are happening in our real world, help us to understand control and avoid them if necessary. But look, if you've been buying your white wines or rosé wines in clear glass for the past five years and you enjoy them and you're still buying them, then go for it. If, on one occasion you do get one that has a weird sort of uncharacteristic smell or flavour that you can't reconcile with your previous experience with it, then it's really useful at this point to understand it's probably due to an avoidable fault. That avoidable fault in that instance is probably not caught, it probably is affected by light strike.
33:43 - Sentient AI: can machines ever be like man?
Sentient AI: can machines ever be like man?
Toby Walsh, University of New South Wales
To the world of artificial intelligence or AI now; Google engineer Blake Lemoine recently captured the world’s attention when he went public with his conviction that the chatbot he was working with, the“Language Model for Dialogue Applications or LaMDA for short, had displayed evidence of having its own feelings and consciousness. Since speaking to the press, Lemoine was put on Google gardening leave, and the company have roundly dismissed his claims. So is this a computer programme with feelings that knows its alive and fears the off switch, or just a clever piece of software? James Tytko asked Toby Walsh, professor of artificial intelligence at the School of Computer Science and Engineering at the University of New South Wales, whether LaMDA could really be sentient…
Toby - No, it's not really sentient, not in my opinion. We don't build machines with anything like sentience. On a strictly technical level, it's not sentient because it's not made of biological stuff and only biological things have consciousness have sentience. But what I think is interesting is how easily even smart people like a senior Google engineer can be taken in. So it says more about human gullibility than it does about intelligence of machines.
James - If it isn't sentient, how is Lambda so good at replicating the speech of a real person?
Toby - Well, it really is auto complete on steroids. It is essentially the same sort of thing that's on your smartphone that can finish the word or maybe finish the sentence when you're typing in a text or an email, but they've taken it to the next level by pretty much pouring in the contents of the internet into this large neural network. So it can't just complete the next word or the next sentence, it can complete whole paragraphs. But it's not understanding what it's saying. It's merely just trying to say things that will frequently turn up on the web.
James - But is it a concern that without him going public with this belief, we would not have known that a chatbot as sophisticated as this was in development at one of the biggest technology companies in the world, Google. Why are they interested in developing chatbots that we find difficult to differentiate from real people?
Toby - Well, there is an arms race going on between the big tech companies to develop these chat bots. And Google's not the only one, that's got a large chat bot; all the other tech giants are developing them. There's nothing special about what Google's done, it's their flavour of chatbot. And they're going to prove very useful. Google is going to be using it to help return better, more accurate search results for you. They're being used in customer service, they're being used in a variety of ways. They can do some remarkable things. They can summarise restaurant reviews for you. They can even write computer code; you just say what you would like the computer code to do and because there's quite a bit of computer code out there on the web that's been poured into this program, it can surprisingly enough actually write computer code.
James - I'm interested in the idea of sentience and AI. And I'm wondering whether, is it even in the realm of science, the idea of sentience, or can we ever expect scientists to understand sentience, to then be able to give it to machines?
Toby - We don't know if it's the stuff purely of biology or it's something that we could reproduce in machines. I always say it will be lucky if machines are never sentient, because then we won't have to worry about them. I can go back to my laboratory and I can take my robot apart, diode by diode, and no one's going to care because it doesn't have any feelings. It's not going to experience any pain. But if it did become sentient, then most things that have sentience, we give rights to like other humans and animals.
James - You've recently released a new book, "Machines Behaving Badly." Can you give me and our listeners a bit of a taste of your rough direction of travel in that new work?
Toby - Yeah. I don't think we have to get into some super intelligent machines - some Terminator-like robot - before we get to have to worry about these things. Indeed, I think we're already there. We already have to worry about the stupid algorithms that are starting to be given responsibilities that are that impacting upon our lives. As an example, the machine learning algorithm used by Facebook to decide your news feed or the machine learning algorithm that's used by Twitter to decide which tweets you read, that doesn't seem to be particularly well-aligned with human values. It's encouraging polarising of our debate, it's encouraging fake news, it's not encouraging a healthy, democratic debate, it's not encouraging us to understand other people's viewpoints.
James - Are there other implementations of AI currently in use in our society that similarly crop up as a point of concern for you?
Toby - Yes, for example, many people don't realise that their boss today is an algorithm. If you work for Uber or Deliveroo or one of these gig economy companies, your work is not decided by a human, it's decided by an algorithm. And it's demonstrated for example, that the algorithm is encouraging bad driving. It's encouraging people to break the speed limit, you'll get more jobs and more money. There's going to be more work for those people.
39:52 - Mendel's 200th Birthday
Mendel's 200th Birthday
Berris Charnley, ARC Centre of Excellence for Plant Success in Nature and Agriculture, University of Queensland
But first, this month marks a special day for the field of genetics: it’s the 200th birthday of Gregor Mendel. Being out in the garden gave Chris Smith pause for thought…
Chris - Now this year, I've upped my home-grown game. And one of the things I'm watering in the hot weather on my pea plants. And it's amazing to think that someone doing something similar growing and tending their peas over 150 years ago, laid the foundations of modern day genetics. The monk, Gregor Mendel, who was born this coming week two centuries ago in what is now the Czech Republic, came from a poor family but he received a very good education. It endowed him with a grasp of maths and statistics that enabled him to spot, track and understand how different characteristics of his pea plants were inherited long before anyone had any idea that DNA even existed. I caught up with science historian Berris Charnley, who told me more about this intriguing man's life...
Berris - Gregor Mendel was an interesting character, and he really has become a larger than life character. He was a monk who was living in Eastern Europe in the mid 19th century. He was in fact schooled in what was, at the time, the most cutting-edge pedagogical educational technique on the continent. And so he had a background in thinking about the world around him and experimenting as a means to find out more about the world around him. So one way of thinking about his work is as an exploration into the idea of whether God made everything all at once, or whether plants and animals do change over time in ways which we can analyse and experimentally recognise. And this was part of what he was aiming to look at in his work with peas.
Chris - What was the concept of genetics before he came along? Because where I grew up, opposite, my house was the home of John Ray, the naturalist who actually went to Cambridge; from being son of a blacksmith, became a scientist at Cambridge university. But I was surprised to read that John Ray was one of the first people to say, well, plants inherit something, meaning baby plants come from a parent plant that looks like they do. And prior to that, people weren't really in tune with the idea that things inherited something that made them look the way they did. So what did people really think about genetics and genes up until that point?
Berris - One school of thought is that when people talked about something being inherited, they were talking by metaphor to legal forms of inheritance. So they weren't discerning between what we might call cultural and biological inheritance in the same ways that we might. One of the key things that changes when people begin to reappraise Mendel's work and think about it more deeply, there's precisely a change in what this word "inheritance" means and ideas about what it is that's being passed from an adult to an offspring, and the biological basis of those ideas.
Chris - Is there evidence that he was familiar with what Darwin had been saying, because Darwin comes along a bit before Mendel - he'd written "On the Origin of Species". Darwin had seen the product of this happening, but didn't know how it was happening. Was Mendel familiar with that and then thought, well, this is some meat on the bones? Or were they completely independent of each other?
Berris - It appears that Mendel was aware of Darwin's work, but the other way around isn't true - Darwin wasn't aware of Mendel's work. The legend is that he has a copy of Mendel's famous paper experiments in plant hybrids and it's in his library. But one can tell that he didn't read it because he hadn't cut the pages. When journals were circulated in the 19th century, they were often bound all the way around; the spine and the outers of the pages were bound and one would have to cut open the pages so you can read them. People have looked at Darwin's library and found that he didn't actually get to read Mendel's work. Darwin had his own theory of inheritance which was quite different to the work which Mendel was doing. Darwin's theory suggested there were things called "pangenes", which circulated within an animal and then passed through to its offspring. What happens in the 20th century, people bring together Darwin and Mendel's work to form what we think of as the basis for our modern ideas about evolution and inheritance.
Chris - Did Mendel have much insight into the difference that his ideas were going to make? Or was he pretty self-effacing and humble about this and thought, "well, I found something interesting here. I've written it up. I've been good to my scientific training and background. Now I'll move on and do other administrative duties." I think one commentator once put it that, that he succumbed to that disease that all academics do, which is that they get consumed by admin. They don't get to do any science anymore. What is the end of the story?
Berris - It's very true that Mendel doesn't really pursue a scientific career in the way that we might think of a scientific career. So he doesn't build on his initial discoveries and follow them through to other discoveries. He does continue doing some experiments with some other plants, but the general feeling is that Mendel feels like he's discovered something of moderate significance to what might be a special case, which doesn't really provide the groundbreaking world-shaking answers to either the question about where new types of sheep might come from, or the question about where new species of animals might come from.
46:03 - Octopus-inspired glove aids underwater grip
Octopus-inspired glove aids underwater grip
Michael Bartlett, Virginia Tech
Have you ever tried to grab something under water? Chances are, you slipped and slid all over the shop trying to get hold of it. Octopuses, on the other hand - or perhaps that should be tentacle! - seem to have no problem sweeping up objects and moving and manipulating them. How they do this was the inspiration for a new wearable piece of tech that Michael Barlett, a mechanical engineer from Virginia Tech, showcased for Julia Ravey…
Michael - A lot of people know that the octopus has eight arms, but it also has over 2000 individual suckers across those eight arms. And those suckers allow it to grab objects and then to quickly release them. What the octopus also has is millions of sensors across those adhesives and those arms, which allow it to actually feel its environment, and chemical sensors actually allow it to taste its environment. So what the octopus is able to do is to take information from those sensors and then process it in its brain. This ability to have adhesion control with the ability to sensor your environment, process that information, and then that allows the octopus to actually individually control those thousands of adhesives to pick up and release objects.
Julia - How does this compare to humans when we're underwater trying to grip things?
Michael - A lot of us can relate to the concept of having to pick up something wet and slippery underwater. It's quite hard. So what the octopus can do is actually with their suckers, or with this control of adhesion, they can pick up things without actually having to grab them. They can just lightly press their skin onto an object and that allows them to attach to that object reliably.
Julia - You've designed a product which makes us a little bit more octopus. What did you come up with?
Michael - Exactly. So what we wanted to do was to develop adhesives to mimic the suckers and take advantage of that concept of using sensing, processing, and control to manipulate objects underwater. So first we needed to make an octopus inspired adhesive and this consists of a rubber stop with a rubber membrane on top. We changed the shape of that membrane to rapidly change adhesion, about 450 times different from an attached state to a release state and we do that in less than 50 milliseconds. We then needed a sensor. So then we used what are called micro-LiDAR sensors, which essentially look at their environment and detect objects that get close to it by shining beams of light. We take the signal from that, we send it through a microprocessor and that allows us to detect objects in realtime underwater. Next we took this sensor and this adhesive and we put it onto a glove. We can now automatically trigger adhesion just by moving your hand next to an object, similar to how the octopus would grasp something underwater.
Julia - So if I had this glove on and I wanted to pick something up, say in my sink, and there's a plate in the sink and there's lots of water in there, and I had this glove on - what would be the process by me putting my hand in the water? How could I pick it up and then release it wearing this glove?
Michael - A plate is a great object to try to think about picking up; you can't simply just put your hand on a plate and pull it out of water. But if you had one of these gloves on, what we call Octaglove, you would essentially be able to take your hand in a flat state, put it on the flat part of the plate, and if you just move your hand close enough, the sensors will detect the plate, which then will rapidly trigger the adhesion to turn on. All of a sudden, you'll be attached to that plate, just like an octopus would be attached to a shell underwater, and then you could pull it out of the sink and then put it in your drying rack.
Julia - It's like a magnet, but not magnetic.
Michael - It is similar to that in the sense that you just have to get close enough, and then the adhesives and the glove take care of the rest to pick up that object.
Julia - What do you think this octopus-inspired technology could be used for then?
Michael - I think some of the areas we're excited about are certainly underwater grasping and moving of objects, you know, in the same way the octopus would. So I think you could imagine undersea divers being able to pick up objects that are delicate. We also envision it being useful for assistive devices and rehabilitation where a user could simply bring their hand close to an object and then the adhesion could turn on to allow them to pick up that object effortlessly. And I also think there's possibilities in manufacturing and other industrial settings.
50:37 - QotW: Would a moon explosion be more dangerous?
QotW: Would a moon explosion be more dangerous?
This is a great question. Sci fi films love to show a good explosion on the moon, but what would actually happen if this was to occur? Well, our great friend and University of Cambridge’s public astronomer Matt Bothwell, is here to make sense of this outer-space occurrence.
When explosions happen on Earth, there are two things that can cause damage: the shrapnel, and the blast wave. Shrapnel are the pieces of metal from the actual thing that got exploded, which get pushed outwards at high speed by the explosion. As you can imagine, getting hit by a piece of shrapnel is pretty bad for you. On the Moon, because there is no air the shrapnel won’t be slowed by air resistance. Plus Lunar gravity is weaker than Earth gravity, so the shrapnel will fly much further and much faster before hitting the ground.
Any shrapnel from a Lunar experiment will be much more dangerous, over a much longer range. This is the same reasoning of why space junk is incredibly dangerous to satellites in our orbit.
The other part of the explosion, the ‘blast wave’, is a pulse of high-pressure air which travels supersonically outwards from the explosion. Blast waves can be very damaging — they can knock buildings down. The good news for us is that there is no air on the Moon, which means no blast wave. The same goes for meteorite collisions. When meteors hit the Earth, one of the most dangerous things is the air burst, when the meteor slams into the Earth’s atmosphere and causes a huge explosion. But on the moon, no atmosphere means no air-burst.
The blast wave we experience on Earth is the air molecules being pushed together and propelled outwards from the explosion. This is usually more deadly than the fireball an explosion would produce. But if there’s no air to compress, the blast wave can’t occur and so the explosion is less dangerous. When we posted this question to our forum user evan_au agreed with Matt that in the vacuum of the moon, shrapnel would travel further and be more dangerous.
So Neil. Our Lunar explosion will have more dangerous shrapnel, but no blast wave at all. What that actually means in terms of how powerful or dangerous our explosion is, is hard to say without an experiment. We just need to convince NASA!
Related Content
- Previous Odds on Either Way...
- Next Suicide
Comments
Add a comment