Shingles vaccine delays dementia, and chatting AI bots

0:11Hello,

0:17welcome to the Naked Scientist podcast, the programme that brings you the biggest breakthroughs and also talks to the major movers and shakers in the worlds of science, technology and medicine. I'm Chris Smith. Coming up, another study confirms that shingles vaccines appear to knock dementia risks on the head. The BBC's Zoe Kleinman explains a social media site just for AI bots to hang out. And will the US or China be the ones to make the next giant leap in the new lunar race.

0:55As we live longer, the number of people who are destined to develop dementia is expected to rise dramatically in the coming decades, which makes it one of the most important public health considerations of our time. New research, however, offers fresh hope. A study which has been published in The Lancet Neurology has found that herpes zoster vaccination which combats a manifestation of the chickenpox virus called shingles and is marketed under the brand name Zostavax, can reduce dementia risk by about 20%. The findings mirror results of an earlier initial study carried out on vaccinees in Wales by Stanford University's Pascal Gelsetzer.

1:36I asked him to explain what they did during that first study, which we also featured here previously on The Naked Scientist, and then to tell us how his new Canadian results compare. We were using a natural experiment in Wales to show a large preventive effect of shingles vaccination for dementia. And what's so unique about this study is that we were able to take advantage of the fact that the NHS, the National Health Service, rolled out the vaccine using very specific date of birth cut-offs such that you were ineligible if you had your 80th birthday just prior to the start date of the programme but you were eligible just after.

2:14And this provided us with these beautiful, clean comparison groups that should be similar in all characteristics to look at dementia as an outcome. And we really overcame in this way the usual confounding bias concerns that we have in such observational data analyses. And correct me if I'm wrong, you saw about a 20 % reduction in the rates of dementia diagnosis in the seven-year follow-up period after the jab was administered in the people who did get it versus those who didn't. Correct, yes. So that's a dramatic reduction in rate of dementia, isn't it? Why do you think you saw that? Absolutely, this would have huge implications for population health, clinical medicine, for dementia research.

3:02It's a one-off, inexpensive, intensive, simple, readily available, safe intervention. There are two broad mechanisms here that we can think about. One of them is that the chickenpox virus, of course, remains hibernated in your nervous system for life after you contract chickenpox. Reactivations of the chickenpox virus with age probably causes some inflammatory process. We know inflammation is a risk for many chronic diseases, including for dementia. And then another pathway could be through indirect effects on the immune system, where we're increasingly understanding that vaccines appear to have broader effects on the immune system beyond that specific antibody response that they have been designed to elicit.

3:46And what did you do this time to corroborate the effect you saw in Wales? We were looking at Canada, primary care data from Canada, which provides a really nice opportunity to corroborate the findings from Wales because they used two different date of birth eligibility cutoffs in their rollout. And that was just used in Ontario. So we can look at whether we see the same effects at both of these date of birth cutoffs. And we can compare if we see effects only in Ontario and not in other Canadian provinces that didn't have a shingles vaccination programme. And in fact, that's exactly what we see.

4:27We see the very same patterns that we see in Wales. We also had a paper out last year where we show, again, the same effects in Australian data. So it really seems to hold up in different data sets, different populations, different health systems, and also for different dementia-related outcomes. And the scale of the difference you got in Canada was equivalent to what you saw in Wales, was it similar yes we saw large protective effects yes so you appear to have nailed this it's not a statistical glitch it does look like this is a real finding you've tested it in multiple geographies on multiple ages and in large groups of people now so what are you going to do next for us i think the crucial next step now to really conclusively test this research question which is such an important one, I think, for population health and dementia research, is to try to run a large-scale clinical trial of shingles vaccination for dementia.

5:27And for that, we would like to ideally use the old life-attenuated shingles vaccine, which is the vaccine that we have used in all our natural experiments or that we've studied in all our natural experiments. Unfortunately, that vaccine is of patent. There's no commercial interest. And so we'd really like to try to raise funds from philanthropy, private foundations. If there's anyone out there who wants us to help get this project off the ground, please, please reach out. Yes, indeed, because we've changed the shingles vaccine that we're using in many places now. We're not using live vaccine anymore.

6:01We've switched to a non-live version. The argument being this is safe for people who can't have live vaccines. Do you think then that that might be less good at preventing dementia? It's an open question. I think it depends on the mechanism that we think is at play. If it's a chickenpox virus specific mechanism, then we would expect the new vaccine to have even greater benefits because it's more efficacious at reducing the reactivations of the virus. If it's more of an indirect immune system mechanism, then it may well be specific to the life attenuated vaccine because we know these mechanisms are particularly strong for life attenuated vaccines.

6:44Fascinating findings aren't they? It's going to be really interesting to see what turns up longer term from this work. That was Stanford University's Pascal Geldsetzer. Now barely a day goes by without another story about AI and its impacts both good and bad. Most people are now very familiar with systems like chat GPT which can summarise large bodies of text or extract information from a broad range of sources, notwithstanding the occasional inclusion of odd bits of made-up information, otherwise known as hallucinations too. But the newest development on the tech scene is what's called agentic AI.

7:19These are systems that can carry out tasks on their own without constant human input. You point them at a problem or a job and let them get on with it. But an interesting phenomenon has been unveiled recently in the form of a social media site exclusively for these bots to meet and exchange information. It's called Maltbook, and what's on there has people both intrigued and spooked in equal measure. We asked the BBC's technology guru, Zoe Kleinman, to explain what's going on. Maltbook claims to be a sort of social network for AI tools to chat amongst themselves, if you like. and you can go to it and have a look at it.

8:01It says humans can read but can't post and it looks if you use Reddit which is a massive chat forum for humans on the internet it looks a lot like Reddit and it's got different threads of conversations if you like but the conversations it claims are all generated by AI tools talking amongst themselves not talking to us. However they They are talking about us. Sounds spooky, doesn't it? How do they get there in the first place? Is this human operators for fun pointing their AI systems at Maltbook? Or are these things genuinely stumbling upon this on the internet? Because that would be more intriguing if they were finding it for themselves and then registering themselves and then starting a conversation for themselves.

8:50Yeah, I mean, I think it's important to be a little bit sceptical about this. You know, we don't actually know how much of this is really genuinely AI generated and how much of it is humans messing about. And when you read the posts, they do come across as being like we would imagine AIs to talk to each other, I think, rather than necessarily how they would communicate. They use very human terms. You know, they're talking about things like consciousness and sentience and even sharing sort of movie recommendations. It kind of feels like what we think AI sounds like rather than perhaps what AI does sound like.

9:29And another thing to bear in mind is it claims to have one and a half million members, which is, you know, AI bots. But there's been quite a lot of scrutiny of that. And one researcher reckons that half a million of them seem to come from the same address, from the same machine, which sort of suggests that something's not quite right there. But the idea is these are agents that are able to access people's material that are helping them out. And we're not talking about sort of chat GPT leaping on there, if you like. We're talking about specific AI tools designed to work with specific people on specific things and using this open source that is signposting them to the site.

10:16If you believe what's said about it. These sorts of systems have been known to communicate with each other in the past, haven't they? Because some of the big social media sites that do advertising and things like that, have they not had these sorts of systems even evolve their own means of communicating, which actually led to people being somewhat spooked and turning them off because it excluded humans from the equation? Yeah, there have been stories of AI tools sort of speaking in a language that's not recognizable to humans. I spoke to one big AI company once that told me it set some up to carry out a test to see how easy they were to hoax.

10:55And so one AI bot was charged with trying to get the other one to give up confidential information. and what they told me was that it really didn't go very well because it failed. The AI tool couldn't get the other one to tell it what it was trying to get it to say and eventually it just sort of gave up and said, oh, you know, I've just been having you on, I've been trying to fool you and then the one that it had been trying to pull said, oh, you were really good, you nearly had me, you were brilliant and they started being like super sycophantic with each other, you know, telling each other how great they were And they sort of had to stop the experiment in the end because it just descended into this kind of quite sickly, you know, AI mutual appreciation society.

11:40But that's one of the other things that I know you as a technology correspondent have almost certainly spoken to publishers. I've been speaking to science publishers. They're very worried because a lot of the content on the Internet is now being generated either by or with the support of these sorts of systems. and a they make mistakes and b they do generate stuff which isn't actually that good and so people are now concerned that we're building up a sort of a stratum in the knowledge base of ai rubbish which in fact is going to pollute the knowledge space but other ai systems of the future will try and learn off that stuff and that may deform and distort things yet further it is worrying isn't I think AI slop, as it's called, is sort of the new internet spam, isn't it?

12:29And when you see, you know, just rubbish, rubbish, and very obviously AI generated content, whether it's a picture or a video or even text, you can spot it pretty quickly because it's bad. And you sort of roll your eyes at it. But you're right, you know, if it's getting sucked into the data that's used to train future AI products, And then is there a quality control, you know, boundary in there? And will anybody be aware that this stuff is nonsense, that, you know, Jesus is not, in fact, half shrimp? I don't know why there's a lot of shrimp Jesus is around, by the way, but there are. You know, and you're absolutely right.

13:11I mean, I suppose you could say in the history of human storytelling, we've always done things like that, haven't we? you know the urban myths we've got um legends dating back centuries we don't really know whether they're true we suspect they're probably not so i guess human history itself and and the things that we learn from isn't 100 reliable either um but i think you're right the thing with the ai slop i suppose is what's it doing is it there to entertain you know if you're watching a i don't know a surfing dog do you care whether it's real if you like it and it's funny and entertains you or does it matter more if it's something that's deliberately designed to be manipulative or deceptive and there is a lot of that as well.

13:57Zoe Kleinman at the BBC. The Naked Scientist podcast is produced in association with Spitfire. Cost-effective voice, internet and IP engineering services for UK businesses. Find out how Spitfire can empower your company at spitfire.co.uk.

14:19Music in the programme is sponsored by Epidemic Sound. Perfect music for audio and video productions. This is The Naked Scientist with me, Chris Smith. Still to come, the US stalls in the race back to the moon. Why is that and will it give China an advantage? But first, to microplastics. which have now been found in fish even in remote parts of the Pacific Ocean. Fish are of course an essential source of food for communities that live on islands like Fiji and Vanuatu but around a third of those studied by researchers at the University of the South Pacific were found to have ingested microplastics.

14:54This might have significant impacts on human health and on the environment and the results also align strongly with findings in other marine environments worldwide. Richard Thompson at the University of Plymouth is considered the world's leading authority on microplastics, and I've been speaking with him. Particularly at this time when we've got the Global Plastic Treaty under negotiations, although 174 countries have agreed that there's a problem with plastic pollution, they can't agree at the moment on how to address that issue. And so having clear evidence that this is a problem that's affecting all nations really helps to galvanize activity and focus activity on solutions.

15:34The quantities found in these fish in the Pacific Islands, I was really interested. They're very similar to what we reported from fish in the English Channel more than a decade ago. You know, we showed about 35 % of the fish were contaminated. We looked at 500 fish. This paper reports 32 % of the fish contaminated. And similarly, it reports a presence of a lot of fibres amongst the types of microplastic that are reported. I think the evidence is key to helping bolster the action that needs to be taken by all nations to reach a consensus on the need for a legally binding treaty to end plastic pollution.

16:13And, you know, for some of these countries, these small island developing states, of course, they're not producers of plastics. they're importing it but in some cases you know don't have any real means of dealing with the associated waste and so it's accumulating and causing a major problem. Are the fish getting their plastic pollution from local sources or is it washing there from western and developing countries that are just chucking it in the sea? That's a really good question and I think you know the paper uses statistics to show differences between some of the different islands. Of course those differences don't tell you about the causality.

16:52And so I would be cautious about the extent to which the paper can infer what the underlying causes to these patterns were. I mean, back in 2014, we published a paper showing very high accumulation of microplastics in Arctic sea ice. And we were reporting greater concentrations there than in the waters close to me in Plymouth, which is obviously a big city. So the key point with particularly the microplastics the particles they can be redistributed by wind and tide and can be reconcentrated at places that are a long way from their original source so i think it's hard to be sure where exactly these plastics and in the pacific islands are coming from but it's clear that they're accumulating just as they are elsewhere on the planet when i read this paper something interesting stood out to me which i hadn't really thought about much before which is that different plastics behave differently as in some of them float more some sink more and therefore different species which live in different ocean domains are differently affected by these different plastics aren't they yes again and it's a question that we we also looked at with the samples we took from the english channel about a decade ago we didn't see a clear effect in in that data between the demersal the bottom feeding fish and the plagic so it's interesting that this paper is revealing some trends.

18:12Of course, the density of the plastic will affect its transport potential. Also, the shape of the plastic will influence its transport potential. And so it doesn't surprise me that then means that some species have a greater likelihood of encountering the plastic than others. And it's interesting to see this data. There's a much bigger sample of fish here, 878 fish examined compared to the 500 or so that we looked at from the English Channel. You mentioned that what we need is a joined up round the world solution to really grab this bull by the horns and do it soon so that we start dealing with this.

18:47But what's the legacy of what's already in the ocean? How long are we going to be living with this problem? Do you have any insights into whether this is a hundred year problem, even a thousand year problem, or if we stop making plastic today, it'll be gone by next week? It's a really key point. And, you know, I wish that I'd got a magic wand that I could wave today that would end plastic pollution entering the environment tomorrow. But even if I'd got that, let's be clear that the quantities of microplastic in the environment would continue to increase for some decades to come because of the fragmentation of larger legacy items that are unfortunately already in the environment.

19:25But that isn't a reason not to take action now. To me, it's an amplification of the urgency to take action. It's a little bit like trying to slow down a large ocean liner at sea. This isn't a problem that we're going to be able to solve overnight, but that's no reason not to put on the brakes. and have we got any clearer pictures yet as to one of the other things that you were kind of instrumental in highlighting which is that it's not just the plastic that's the threat it's the stuff that sticks to the plastic other oil-based chemicals that share chemical characteristics with plastic which means the plastic goes around picking up a cargo of this stuff and can then impart that into the things that eat the plastic yes we know that there's the potential for particle toxicity from the presence of the small particles there's also the issue of chemical toxicity and that can come from chemicals that are required once the plastic is in the environment but of course it can also come from the additive chemicals that are incorporated into plastics during production and in many cases the concentrations of some of those additive chemicals are much higher than are likely to be accumulated from the ocean a recent report suggested there were some 16 ,000 chemicals associated with plastic production, around 4 ,000 of which were known to be potentially harmful.

20:43So I think as we move towards a plastic treaty, it's going to be really important that to make plastic products safer, we consider the chemistry of them and reduce the suite of chemicals that are used and focus on chemicals that are going to be less harmful. Richard Thompson at the University of Plymouth and the study that Richard was commenting on for us there was published recently in PLOS One. NASA had been forced to delay their mission to launch astronauts around the moon by at least a month. It follows a hydrogen leak which the US Space Agency has reportedly encountered previously and had three years to fix.

21:22This is fuelling speculation that the hold-ups could hand victory to China in the race to be the first back on the moon. David Whitehouse is a space scientist and an author and has the latest. It's a refuelling problem and as you said it's the same refuelling problem they had with the uncrewed Artemis 1 launch three years ago. It's the hydrogen feed into the fuel tanks. Hydrogen is a small very energetic molecule that's the reason it's rocket fuel and you don't want any of this seeping around or leaking during launch. And the fact that they still have this problem three years after they had to solve it for the first mission is very disappointing.

22:03And at a post-cancellation of what's called the wet dress rehearsal, when it was realised they'd have to wait till March to send the astronauts to the moon, there were some very uncomfortable engineers in front of the press trying to explain why they still had this problem. Yeah, well, what did they say? why have they still got the problem? If they know that this is the fault, surely it's relatively easy now you know the fault to identify what to do about it. That's a very good question, which I don't think was satisfactorily answered. They said during the launch that had they instituted a few straightforward fixes in time, they would have met the launch criteria.

22:43So they said they could have met that, but they weren't sure and they didn't have time. It's not very satisfactory. There's not a great explanation. They're trying all sorts of fixes in the hope that it will work and that they'll get these astronauts to the moon next month. But it is not, frankly, a good look. Why have they had to delay till March? Why can't they just fix it and then say, right, we're going next week? It's the orbital window. They have to have the moon in the right place for this to work. So the launch windows are only a few days every month or so. So they have to build up a campaign to that launch window.

23:18They're going to launch into a low Earth orbit, which takes 90 minutes around the Earth. And then after that, they fire the rocket, which takes them up into an incredibly high elliptical orbit. They do that once, and that would be spectacular. But when they're at the top of that orbit, they give themselves a little nudge to go towards the Moon. And if the Moon's in the right place, they can swing around it and then come back to the Earth. So technically they're not in orbit of the Moon but they're using the Moon's gravity for a wide loop becoming the most distant astronauts from the Earth ever.

23:55It should be spectacular but everything has to line up, rocket and the Moon included. So what are China doing at the same time that means they're snapping at NASA's tails? They've said they want to go back to the Moon. They say they want to put people on the Moon. They've got astronauts in training for this and they've got a substantial amount of equipment which they are testing. And we see them testing lunar landers. We see them testing the rocket that might work. We see them becoming increasingly adept at orbital maneuvers and having considerable experience of people working in space. They have said they want to go back to the moon by 2030.

24:30The rumor is that they're bringing that back to 2028 because the earliest America can get to the moon is now 2028, and there are also people who think that they're not even going to make that because their system is so multifaceted and complicated that China's simple approach, particularly since they're not going to land at the South Pole, I hear, gives them a lot more straightforward project where they don't have to bring so many different things together. If you had to bet, I'd actually put my money on China at the moment. Where are they going to land then? Because I thought they were all going for the South Pole because that's where the water is and there are various strategic advantages to being down there i didn't realize they were not going there well this is full of rumor we're expecting an announcement sometime soon but you're quite right the initial plans for america and china were to go to the south pole which is where the water is and which could be the start of a build-up of a lunar base there good reasons to go there you also get access to sunlight most of of the month so you get solar power as well, which you don't get near the equator.

25:41The problem is that we found from trying to land unmanned probe there over the last few years, it's more difficult to do so than we realized, possibly because of multiple lines of communication that form as the thing nears the surface, because it's very much on the edge of the moon, and you get reflections and problems with the propagation of the radio signal from the lander, which make things rather difficult. and many probes have crashed there when we didn't expect them to. So the rumour is that for the first landing, China may well take the simplest route and go near the equator. But we haven't seen them do visible displays of space muscle flexing though, have we?

26:25We haven't seen them posturing and saying they're going to do this in a dramatic way. Are they basically keeping their powder dry and their cards close to their chest so they can basically over-deliver having under-promised? Well, this is the way they do things. We've seen very impressive tests of the lunar lander. We've seen them increasingly master orbital mechanics with their spy satellites. They're doing some very sophisticated things. They have a great deal of experience now with people living and working on the Tiangong Space Station. Look at their recent Changi mission to the far side of the moon where communication is very difficult, not only took instruments from India and from two European nations, it overflew the landing site, it tilted to have a sideways look, watched itself go down, it used two different methods of measuring the height above the lunar surface, it touched down very well indeed, released a rover which made some wonderful mineralogical observations of the Moon, and that was a spectacular mission, like of which the West has not performed on the Moon for decades and decades and decades.

27:34So they have the capability. What they don't play is the NASA game of publicity. They seem to prefer just to do it rather than have the great big build-up, which was actually seen with Artemis too. If you can't deliver, it's a bit of a damp squib. Space scientist and author David Whitehouse there. Let's hope next month is a bit more successful. well that's where we leave it for today but do tune in on tuesday when you can catch the buzz all about the biology of mosquitoes how did the world's most dangerous animal evolve in the first place and why how do they track us down and how do dread diseases like malaria and dengue exploit them moreover what can we do to fight back meanwhile thanks to all of you who are supporting us with donations this really helps and if you would like to make a contribution to keep the show on the road, do please head over to nakedscientist.com forward slash donate.

28:32We really, really are grateful. Please also, if you can follow us on LinkedIn and on Instagram. And we are really grateful too, for reviews on Spotify, Apple, or wherever you get your podcasts. We really relish that feedback. I'm Chris Smith. Thank you for listening. And until next time, from all of us here at the Naked Scientist team, goodbye.