Twitter suspended 70,000 accounts after the Capitol riots and it curbed misinformation

Nature Podcast
The Nature podcast is supported by Nature Plus, a flexible monthly subscription that grants immediate online access to the science journal Nature and over 50 other journals from the Nature portfolio.

More information at go dot nature.com plus.

UnitedHealthcare
One size fits all seems like a good idea for clothes until you try them on. Same goes for healthcare. That's why UnitedHealthcare offers flexible, budget friendly coverage for medical, vision, dental, and more. Learn more at, uh one.com.

Unknown
In an experiment we don't know yet.

Unknown
Why is blight so far like it sounds so simple?

Unknown
They had no idea, but now the data.

Unknown
I find this not only refreshing, but at some level, astounding.

Nature Welcome
Nature Nature welcome back to the Nature podcast. This week, creating a new kind of.

Benjamin Thompson and Lizzie Gibney
Quantum matter and how deplatforming users after the US Capitol riot stemmed misinformation on Twitter im Benjamin Thompson and Im Lizzie Gibney.

Nature Welcome
The quantum world is pretty mysterious.

Not only do quantum rules seem so weird, but everything happens on a miniscule scale.

One way to study this world is to make it balloon in size using a bose Einstein condensate.

These condensates happen when atoms get really, really cold.

If you remember the Heisenberg uncertainty principle, you'll know that the more we know about a particle's momentum, the less we know about where it is. So if you cool down a bunch of atoms to close to absolute zero, what we know about their momentum becomes precise. But the boundaries of where the atoms could be get so big and blurry that they eventually overlap. You end up with indistinguishable atoms, all sitting in the lowest energy state, acting like one giant atom, which follows quantum laws but is visible at a macroscopic scale and is precisely controllable.

These Bose Einstein condensates, sometimes called beCs, can be used to study weird phenomena and create quantum technologies. But so far, physicists have mainly made them from atoms. They long to make them from molecules, stable molecules, and ones that interact, which would massively expand the kinds of studies they can do now. Physicists have done that for the first time. I spoke to Sebastian Ville, a physicist on the team at Columbia University in New York, and I asked him what researchers can already do with Bose Einstein condensates.

Unknown
Bose Einstein condensates have been used to study quantum phenomena on a macroscopic scale. For example, it was possible to study superfluidity extremely well, and also in the early two thousands, atomic Bose Einstein condensates opened the field of quantum simulation. For example, efforts have been made to study superconductivity in great detail using these systems. And again, we benefit a lot from the fact that everything is on an atomic scale. In these systems, instead of an electron scale, electrons are a lot smaller. And this increased size or these increased length scales help us to really look at the system in much more detail and try to understand it both better experimentally and theoretically.

Nature Welcome
So why do we then want to do this with not just atoms, but with molecules? How does that make this quantum playground even better?

Unknown
You know, atoms are beautiful. You can cool them very well. But it also turns out that the interactions between atoms are fairly simple. Basically, atoms at these cold temperatures only interact when they are really sitting on top of each other. That is beautiful and straightforward to model theoretically.

But in real life, many systems, many body, many particle systems interact via long range interactions. So, for example, magnets interact at a distance. There is a sort of invisible force between them, electromagnetic force between them, or also in materials, electrons interact with ionic cores of the crystal lattice in which they are moving, which is a long range interaction. So long range interactions are actually everywhere. But in cold atomic decs, they were not there. And one way to actually introduce long range interactions into a system would be to cool molecules and, in particular, polar molecules. Polar molecules are molecules which have a plus charge on one end and a minus charge on the other end. So there is an electromagnetic interaction between the molecules, and this interaction happens at long range. And if we were able to bother condense molecules, we could really tap into this potential of realizing fully quantum many body systems with long range interactions that we can use to simulate more interesting quantum phenomena. That was the dream already in the late nineties, early two thousands.

Nature Welcome
Seems like there's a really rich array of things that you'd be able to study if you had these super cold molecules in a Bose Einstein condensate.

So what's so hard about creating this state for molecules?

Unknown
There is a saying in atomic and molecular physics, which is a diatomic molecule. A molecule with two atoms is one atom too many. The internal structure of a molecule that's made out of two atoms is already so much more complicated. And when you make clouds of molecules, the molecules collide and stick together upon colliding and get lost from the sample. So, in a cooling process, where you always need collisions between particles, so the temperature gets uniform in the sample, you actually start losing molecules, and you lose the molecules faster than you can cool. And in the end, it was impossible to approach even temperatures that would lead to bose Einstein condensation.

Nature Welcome
And what kind of temperatures do you need? How cold are we talking about here?

Unknown
We are talking about very close to absolute zero. It turned out now, in our work, that to get to a Bose Einstein condensate needed to cool to below ten nanokelvin. So that's just, you know, what is it?

Billions of a degree above absolute zero.

Nature Welcome
That is very, very cold. And so how did you manage to overcome that problem then? You needed to stop these collisions happening that created these reactions between the molecules.

Unknown
So there we borrow tricks that are possible once you are using molecules. We all know a very popular polar molecule, and that is the water molecule. We use microwaves very often in connection to water, basically, in our cup of coffee that we want to heat, we shine in intense microwave fields, and they set the water molecules into rotation, which in this case, heat the coffee. So we can do this in a somewhat more precise and more sophisticated way. In our case, we are setting our molecules, in our case, it's sodium cesium molecules, into rotation by exposing them to a microwave field that rotates and also expose them to a microwave field that oscillates. So this combination set the molecules rotation in a very particular way, such that whenever two molecules collide, they start repelling each other. And this prevented the lossy collisions that have been a challenge to the field really for the last two decades. And really, that was the key to enable efficient cooling via evaporation. Via evaporative cooling.

Nature Welcome
And how does that cooling work? What is evaporative cooling?

Unknown
In our sample, we have a way how to selectively remove the hottest molecules in the sample, and we do this evaporative cooling process, and what remains is something much colder. And first you may think, gosh, you need to throw away all your stuff to get colder. And are you left with anything? Well, that is absolutely true. And that's why you need these collisions between the molecules to be not lossy, because if you have these additional losses, you will never end up with a Bose Einstein condensate. So preventing these collisions, these chemical reactions between the molecules, is crucial to be left with cold molecules at the very end of evaporative cooling.

Nature Welcome
And how many molecules were you left with at the end of this process?

Unknown
So we saw the onset of Bose Einstein condensation when we had a cloud of 2000 molecules and fewer Bose Einstein condensates, we got with 200 molecules. And this we can image. Yeah.

Nature Welcome
And so that's something that just wasn't possible before, to make this kind of condensate out of molecules.

Unknown
Yeah. And this was the first time this was observed.

Nature Welcome
So do you have any plans now of what you think you might like to do? What ways you might like to explore this system or the kinds of other quantum system you might simulate. It sounds like there are lots of options.

Unknown
Basically, we now have a new system that really opens the door to a lot of new explorations. We would really like to leverage these long range interactions, interactions between the molecules. And in analogy with water, we would actually be very excited to study crystallization in the quantum regime. Also, we expect the formation of so called quantum droplets, where actually the droplets, in contrast to water, would be a superfluid. But there may also be technical applications that may come now into range, because a Boser Einstein condensate is sort of a pristine form of matter where we know what each molecule is doing it. Each molecule is actually completely identical in the same quantum state. So that is also an entry point to really get full single molecule control.

And with single molecules, there have been already theoretical proposals for 20 years that you could actually use single molecules as qubits in quantum computing. And before I said diatomic molecules are already one atom too many. But once you are actually able to control the internal quantum states of a molecule, it turns out that there are internal quantum states that are extremely long lived, and that's exactly what you need for good quantum computing systems. And that would be another really exciting direction to explore.

Nature Welcome
That was Sebastian Ville from Columbia University in the US. Head over to the show notes for a link where you can find his paper.

Benjamin Thompson and Lizzie Gibney
2024 is a huge year for democracy, with around 4 billion people around the world able to vote in major elections. But the threat posed by the sharing of incorrect information, misinformation, looms large.

These falsehoods, quickly and easily shared on social media, have been described as a serious threat to democratic integrity. But there are huge gaps in what researchers understand of how and why misinformation spreads online.

This week, nature is publishing a series of articles dedicated to those topics.

One research paper looks at whether interventions by social media companies can actually stem the tide of misinformation.

And to find out, the team investigated the events around a particularly troubling moment in recent us history. Adam Levy reports.

Unknown
These are the sounds of the US Capitol on January 6, 2021, a day when us american politics changed forever.

It was the aftermath of the 2020 election, an election that Donald Trump had lost to Joe Biden. But Trump had been vehemently and without evidence, arguing that the election had been stolen. On January 6, this rhetoric turned to violent action as Trumps supporters stormed the Capitol in an attempt to prevent the certification of the election results.

Chuck Schumer
I have never lived through or even imagined an experience like the one we have just witnessed in this Capitol.

Unknown
This is Chuck Schumer, leader of the Senate democratic caucus, commenting on the events of January 6. Later that same day, after the insurrection.

Chuck Schumer
Had been overcome, this temple to democracy was desecrated.

Its windows smashed, our offices vandalized.

The world saw Americans elected officials hurriedly ushered out because they were in harm's way.

Unknown
And for political scientist David Lazar of Northeastern University in the United States, these events had profound implications well beyond the day in question.

David Lazar
So that gets to the very core of what democracy is about. If people don't believe election results and you can't really have a legitimate transfer of power, really, as an american citizen, I was shocked by the events of the day as they were unfolding.

Unknown
And while many factors and events led up to January 6, a lot of attention turned to social media, and in particular, Twitter. Although less widely used than some other social media, Twitter had established itself as having something of a town square function, with many politicians and journalists using the platform. And researchers like David saw Twitter as playing a pivotal role in building momentum towards the attack on the Capitol in the way that it was used both by Trumps supporters and by Trump himself.

David Lazar
He had called for a rally on January 6. He used the language in that tweet will be wild.

Unknown
A major cause for concern was the sharing of misinformation on Twitter, in particular around the validity of the election results.

And so, in the days after the storming of the capital, Twitter took active steps to combat misinformation spreading on the platform.

David Lazar
It eventually culminated in a decision to deplatform a fairly large number of accounts, I think around 70,000 accounts, that were central in sharing election misinformation. You know, the question I was asking myself was, how big an effect did it really have?

Unknown
These actions at Twitter gave David and his team a unique opportunity to peer behind the curtain and attempt to answer this question.

David Lazar
You know, what platforms do is often pretty opaque. They're doing stuff, but you don't know what buttons they're pressing, and you don't know what days they're pressing them on. In this case, we knew what days, and we knew what button.

Unknown
To find out what effect pushing the button had, David and the team analyzed the activity of over half a million Twitter users during the election period. This showed that misinformation sharing had indeed followed an unusual pattern during the 2020 election cycle, in contrast to the 2016 election, it remained high even after Americans went to the polls, as claims that the election had been stolen circulated within the group of users the team investigated, just over 1000 were deplatformed following January 6, while a small number, they were responsible for almost a quarter of the misinformation links shared in the total sample group. The previous month, Twitters sudden deplatforming of a small group of major misinformation sharers was followed by a dramatic drop in misinformation on the platform. In part, this was due to those spreaders no longer being able to tweet. But that wasn't all.

David Lazar
That had a secondary effect, that the accounts that followed those deplatformed accounts shared less misinformation. And then the third effect was that a lot of the people who behaved a lot like the deplatformed accounts left, and one can guess that their inference was that Twitter was no longer a friendly platform to them.

Unknown
Overall, these three processes saw the share of misinformation roughly halved among the sample group. But how can we be confident that these effects were caused by Twitter's moderation decision?

After all, even at the best of times, there are confounding factors, and the period following January 6 was, by many definitions, not the best of times. And so the authors tried to unpick why followers of deplatformed accounts ended up spreading less misinformation.

David Lazar
They did share less misinformation, but that could have just been the events after January 6, right? Everyone was shocked, and maybe they said, oh, you know, I shouldn't share stuff like this. So what we did was we compared those individuals who followed the deplatformed accounts to other individuals who did not follow those accounts. And what we saw was, for the first group that followed those deplatformed accounts, the amount of misinformation that they shared dropped quite a lot relative to the accounts that did not follow the deplatformed accounts, which suggests that the change we observe is indeed causal.

Unknown
For Emily Thorson, a political scientist from Syracuse University in the United States, it's important to be cautious about how strong a causal link we can draw between the deplatforming and the drop in missing information.

Still, though the results could help guide decisions that social media platforms make in the future, this does give us some.

Emily Thorson
Evidence that if platforms wanted to, they could decrease the amount of misinformation or other types of harmful content on their sites.

Unknown
Emily, who didn't work on this study, although she does collaborate with David, feels that there are still huge unanswered questions on how social media interventions work, or could work in the future.

Emily Thorson
I think it would be potentially more effective to look at who are the people who are either spreading or seeing lots of misinformation and how do we target interventions at them. So who are these people and what's driving them, right? What are the factors, the offline factors that are pushing them to share more misinformation, whether that's media exposure elsewhere, like on television or radio, whether thats interpersonal contact, etcetera.

Unknown
Whatever we learn about the spread of misinformation on social media, our societies still need to decide what to do with that information.

For now, in many contexts, decisions are left up to the management of social media companies.

And since the events investigated by this study, twitters processes have changed dramatically. Now overseen by Elon Musk, the platform has rebranded as x and has dramatically scaled back moderation efforts.

How to strike the balance between protecting users and safeguarding free speech remains a topic of fierce debate.

David Lazar
The question really should be, where does society step in? And, you know, that's a hard one, and I think this is one that modern policymakers are wrestling with.

Unknown
Of course, different policymakers will come to different conclusions in different contexts. But Emily warns that we shouldn't assume that social media is the only or even the primary way that people encounter misinformation.

Emily Thorson
It is still the case that most of the information that people get about politics does not come through social media. It is coming through television, it is coming through radio. It is coming through interpersonal conversation.

Unknown
This chimes with the view that overall exposure to and sharing of misinformation online may be overstated. For example, in this work, only 7.5% of users shared any misinformation during the study period.

Even so, social media does still play an important role in our lives, a role that is constantly evolving. And research like this could help sculpt the decisions we make about what our social media looks like in the future.

David Lazar
I am still concerned about the future that we haven't, as a society really figured out how to manage the whole content moderation. It is a legitimately difficult question, but I don't think we've come up with an answer that protects us going forward.

Benjamin Thompson and Lizzie Gibney
That was David Lazar from North Eastern University.

You also heard from Emily Thorson from Syracuse University.

To read David's paper and more nature articles investigating misinformation, head over to the show notes for some links.

Nature Welcome
Finally on the show, it's time for the briefing chat, where we discuss a couple of articles from the nature briefing. Benjamin, what have you been reading this week?

Benjamin Thompson and Lizzie Gibney
Well, Lizzie, I've got a story about a new antibiotic. Okay, and it's one that kills deadly bacteria but potentially spares the gut microbiome. And I read about it in nature.

Nature Welcome
That sounds really important because that's a massive problem. Right. With antibiotics is they do a great job when you need them, but they end up destroying your microbiome.

Benjamin Thompson and Lizzie Gibney
Absolutely right. So they are a medical marvel, antibiotics. But yeah, on the podcast a few weeks ago, we heard someone describe the gut microbiome as like a jungle. And sadly, antibiotics are kind of indiscriminate, kind of a forest fire. And that can really upset the gut microbiome that's so important for so many different things, for digesting food and for the integrity and health of organs and even potentially for the health of immediate offspring. Right. So, really, really important.

Nature Welcome
I find we have like a story every week on something new about the microbiome and how important it is. So we might only know half what's essential.

Benjamin Thompson and Lizzie Gibney
The list is getting longer and longer. But what happens then? Say we need to treat pathogenic bacteria, bad bacteria, okay? We do that with antibiotics. And in particular, there's this one broad group of bacteria called gram negatives. Okay? Gram negs, they're sometimes called. Within this group, there are some really nasty bacteria, your E. Coli, your klebsiellas, your pseudomonads, and some of these are resistant to common antibiotics. And very troublingly, some are resistant to multiple antibiotics. So we really need to work out new ways to treat these bacteria. And gram negatives are particularly tricky, and that's where this new work comes in.

Nature Welcome
Ok, so what is it that they've done or found?

Benjamin Thompson and Lizzie Gibney
Well, in this particular case, they've developed an antibiotic that they call lollamycin. Okay? Now, lol doesn't mean laugh out loud. In this case, it means localization of lipoproteins. Okay? Now, the lol pathway is really, really important to gram negative bacteria, okay? It's what they use to put these things called lipoproteins in their outer cell membrane. Okay? So without it, they can't make their membrane properly. And so what lollamycin does is it kind of selectively targets this pathway, and it seems to be quite effective at killing inverted bad bacteria.

Nature Welcome
So in that way, it targets just the bacteria we want it to target broadly?

Benjamin Thompson and Lizzie Gibney
Yeah, that's right. So they tested it against well over 130 multidrug resistant strains in a dish, and it seemed to kill those. And in mice that developed bloodstream infection after exposure to these antibiotic resistant bacteria, all these mice survived after being given lolomycin, while 87% of those who didn't die within three days. So quite effective. But in mice, double underlying at this point. And also, lolomycin treatment didn't seem to cause any observable changes in the gut microbiome and spared mice from c. Difficile infection. And this is, of course, an opportunistic pathogen that sometimes take hold in the guts after antibiotics are used.

Nature Welcome
Wow. So I guess we still need to figure out if there are other side effects. There's a long journey to it being used clinically.

Benjamin Thompson and Lizzie Gibney
Oh, 100%. And I think the gap between mice and humans is a pretty big one. And of course, it's a long road to understand exactly how effective it is if any resistance arises. But I think it's worth saying there is a desperate need for new antibiotics, particularly ones that target these gram negatives. Okay. Now, the golden era of antibiotic discovery was in, like, the forties and fifties and sixties. And since then, recently, it's really sort of a trickle of new ones that have come through. And partially that's because it's quite tough to bring these things to market. And also there's not a huge amount of incentive for companies because resistance to antibiotics develops. So if your drug becomes ineffective within a couple of years, there's no point spending countless millions of dollars. So I think this is an interesting one. Like, it's really important that we have these candidates and there are others that are sort of working their way through the preclinical studies. And so it remains to be seen how far this one will go. But say, you know, potentially a good news story.

Nature Welcome
Well, it's nice to have a good news story from time to time. I think I've got what is kind of a good news story, an interesting news story. Mine is also on biology, so I'm well outside of my comfort zone here. This is about discovering. This is a record breaker, a very tiny record breaker. The organism with the biggest genome ever discovered.

Benjamin Thompson and Lizzie Gibney
All right. Okay. So when we talk about genomes, then we're talking about its genetic a's, t's, G's and C's.

I suppose we've talked about bacteria there. They've got pretty small genomes, humans have got a pretty big one. But what's the record holder then?

Nature Welcome
It's a plant, which, I mean, even just that kind of astounded me in the first place. So we're talking about 160 billion base pairs. So that's bigger than the human genome. Do you know how much bigger?

Benjamin Thompson and Lizzie Gibney
Oh, I'm going to say that's got.

Nature Welcome
To be an order of magnitude bigger, 50 times bigger.

Benjamin Thompson and Lizzie Gibney
Right.

Nature Welcome
So that's a lot bigger. So this is also a record. So it's 11 billion more base pairs even than the previous record holder, which was also another plant, which. No dissing plants, but I haven't even written down that plant's name. But it's really fascinating to study because, of course, this plant, you know, it looks, if people know Rosemary, it looks like a little spring of rosemary, but it's actually native to New Caledonia and surrounding areas in the south Pacific. But it's just a little fern, and only a very small proportion of its DNA is actually used to code proteins. And so there are a couple of big questions that scientists have. How does it find the bits it needs when it's got a rifle through 160 billion base pairs?

Benjamin Thompson and Lizzie Gibney
Right.

Nature Welcome
And how has it ended up with so many if it doesn't use them normally? It takes a lot of resources, it takes effort to keep replicating your genome.

And why does it do it if it doesn't need it?

Benjamin Thompson and Lizzie Gibney
I guess in a bunch of organisms there is some redundancy in terms of genes and pseudo genes and what have you, but this looks like there is an awful lot of it.

Nature Welcome
Absolutely loads. And so the answer is, we don't know yet, but there's some speculation and the answer is great. But also, as Max Kozlov, my colleague here at nature, wrote in this story, it might be quite a boring answer, which is that it might be that it's just not very detrimental to this particular plant because it maybe doesn't have a lot of competition. It's in this quite stable environment. And so maybe it can just get by replicating this enormous genome, because it doesn't really have that much pressure on its resources in that way. So it's just accumulated over time and it hasn't had the evolutionary pressure to get rid of it.

Benjamin Thompson and Lizzie Gibney
So we have this record breaker then. Is this just one of these, like. Well, that's a strange thing. Okay, cool, let's move on. Or is there more to be done with this genome?

Nature Welcome
Well, there is definitely more to be done. So I think one thing is they would love to sequence it.

Benjamin Thompson and Lizzie Gibney
So, I mean, I guess they know how long it is then, but they actually want to know the exact order of where the different bases are. And that's gonna be a big job. Right, because that's a big genome.

Nature Welcome
Well, that's it. You've got 160 billion of these things, and it just sounds like an enormous puzzle. So current methods just aren't up to scratch? We think so. You know, finding enormous genome like this helps us to think, well, maybe we should be finding better ways to sequence. And also, researchers want to understand, does having an enormous genome like this have an effect on organisms? Does it help it grow, flourish, deal with climate change? Maybe be resilient? That's independent of the actual DNA sequence. It raises a lot of questions and just sounds like a fascinating, if teeny tiny plant to study.

Benjamin Thompson and Lizzie Gibney
Well, always lovely to have a record breaker on the show. And congratulations to that plant, I suppose. Let's leave it there for this week's briefing chat and listeners. For more on those stories and for where, you can sign up to the nature briefing to get more like them delivered directly to your inbox. Check out the show notes for some links.

Nature Welcome
And that's all for this week. As always, keep in touch with us on x. We're ature podcast or send an email to podcastature.com. i'm Lizzie Gibney.

Benjamin Thompson and Lizzie Gibney
And I'm Benjamin Thompson. Thanks for listening.

Nature Podcast
Deep dive into the world of science with nature. Plus, from the vastness of the distant star systems to the intricacies of infectious diseases due to climate change, we've got you covered. Enjoy access to over 55 cutting edge journals, breaking scientific news, and over a thousand new articles every month. Whether you're a seasoned researcher or just curious, NaturePlus simplifies complex studies. Plus, it's all available right at your fingertips on nature.com naturePlus, the key to unlocking the world's most significant scientific advances. Subscribe today at go dot nature.com. plus.

UnitedHealthcare
What does it take to drive innovation in today's AI driven world? Join executives from across industries as they dive deep into how technological advancements are redefining leadership from generative AI to cybersecurity, cloud computing to sustainability. The conversations with Leaders podcast from Amazon Web Services covers it all. Join the conversation and subscribe today. Available on all major podcast platforms.

Twitter suspended 70,000 accounts after the Capitol riots and it curbed misinformation

Primary Topic

Episode Summary

Main Takeaways

Episode Chapters

1: Introduction

2: The Impact of Account Suspensions

3: Broader Implications for Social Media

Actionable Advice

About This Episode

People

Companies

Books

Guest Name(s):

Content Warnings:

Transcript