Don’t Trust What You See On TikTok This Election Season
Midterm elections in the United States are just under two weeks away. And new research suggests a significant risk of misinformation for American social media users—particularly from the video-sharing platform TikTok. Cybersecurity researchers at NYU published their findings after submitting misleading advertisements to YouTube, Facebook, and TikTok.
The ads contained either the wrong dates or voter requirements for upcoming elections, or perpetuated narratives about the validity of past elections. And while TikTok prohibits all political advertising, 90% of those test ads were approved. Meanwhile, YouTube performed the best in rejecting all of the ads, and Facebook accepted about 30% of English-language ads.
New Scientist’s Tim Revell joins co-host Kathleen Davis to talk about the misinformation implications of social media advertisements. Plus, the dramatic electrical charge of swarming honeybees, the good news about declining monkeypox cases, and other stories.
Invest in quality science journalism by making a donation to Science Friday.
Tim Revell is Deputy United States Editor for New Scientist in New York, New York.
KATHLEEN DAVIS: This is Science Friday. I’m Kathleen Davis.
KATHERINE WU: And I’m Katherine Wu. I’m a staff writer at The Atlantic based in Connecticut. I also have a PhD in microbiology.
KATHLEEN DAVIS: And despite our names being very similar, we are actually two different people.
KATHERINE WU: Can confirm. Later in the hour, we’ll be talking about sars-cov-2 variants and what we Can expect from our third full winter with COVID-19. And we’ll have a conversation with Braiding Sweetgrass author, Robin Wall Kimmerer.
KATHLEEN DAVIS: But first, midterm elections are almost here in the US. And with high political stakes in races all over the country, you’d hope that voters would be equipped with the best possible information at the polls. New research from a cybersecurity team at New York University suggests you should take social media ads with a grain of salt, though, in particular, ads found on TikTok, which they say had the worst record for catching misinformation in advertisements. Joining me to talk about this story and more is Tim Revell, Deputy US Editor for New Scientist. He joins me from New York. Welcome back, Tim. Great to have you here.
TIM REVELL: Thanks for having me.
KATHLEEN DAVIS: Tim, I use TikTok sometimes. It is something that I use to send silly videos to my friends. But it seems like there’s a more nefarious side here, especially when it comes to the ads. Can you walk me through this story?
TIM REVELL: Yeah, so I use TikTok in exactly the same way. And as you flick through, you’ll see that occasionally, ads show up in your feed as well. And so this team at New York University, they looked at how easy it is to get election-related disinformation onto the platform. They actually also looked at Facebook and YouTube so that they could do a comparison.
And to do this, the team created election-related ads that all included either some sort of false election information, so something as simple as just including the wrong date for the election, or information designed to discredit the election process, so such as calling into question the validity of mail-in votes. And what they found was that all of their ads they knew violated the policies of YouTube, Facebook, and TikTok. But YouTube was the only place that managed to reject all of their ads.
Facebook accepted 30% of the English ones and 20% of the Spanish ones. But then TikTok accepted a whopping 90% of them, both in English and in Spanish. And what they did was they didn’t actually put any of these ads live. Instead, they just tested whether the platforms would accept them. And then they never actually got to the stage where anyone would have seen them.
KATHLEEN DAVIS: So just yesterday, billionaire Elon Musk bought Twitter, which is another social media platform where disinformation has been a problem. Are there solutions to this kind of social media misinformation? And are any of them doable before, let’s say, November 8?
TIM REVELL: I think the chances of any big change before November 8 are very slim indeed. But the researchers behind this work, they do mention two things that could drastically improve the situation. And the first one is that the platforms just need to better enforce the policies that already exist. But at the moment, the enforcement is not very good. There’s either not enough effort, or the systems in place are not good enough to catch them. But the second thing is an idea that’s been around for a while now. And that’s universal ad transparency.
At the moment, most of the platforms release some data about the political ads that appear across their websites and apps. But a lot of that data is not really good enough to get a real time picture. For a start, it’s a bit slow. And second of all, it often doesn’t include who’s really being targeted, which is a really key part of this picture. So even researching what misinformation is out there, who’s being targeted, and what effect it has, is still a really difficult task.
KATHLEEN DAVIS: OK, moving on, there is some buzz about honey bees this week. And by buzz, I mean the low hum of an electric current. Tim, walk me through this story
TIM REVELL: Yeah, this is a really fun story from my colleague Corryn Wetzel at New Scientist. And it’s a completely surprised discovery that honeybee swarms generate more electricity per inch than a storm cloud.
KATHLEEN DAVIS: Wow.
TIM REVELL: And the way researchers found this was they were actually trying to track the weather at a monitoring station near their university in Bristol, in England. And they suddenly saw that there was a spike in their instruments for measuring atmospheric electric charge. And normally, especially if you’re a weather monitorer, that would mean that there was some storm activity brewing. But somewhat shockingly for a day in England, it was actually clear that day.
What they saw was they saw some Western honeybees nearby swarming. And this sort of tipped them off that maybe they were responsible for this spike in electric activity. They managed to record three swarms of bees for around three minutes each. And they found that a bee swarm can generate an electric charge of between 100 to 1,000 volts per meter, which, for that amount of density, is around eight times higher charge than a thunderstorm cloud.
KATHLEEN DAVIS: Now, do bees at rest generate the same electricity? Or is it only when they are flying around and flapping their little wings?
TIM REVELL: So bees, when they move, they do actually generate a little bit of electricity from little hairs on their bodies that rub against other parts of their body. But most of this electricity we think probably comes from when they’re flapping their wings. And that has friction with the air. And that creates a sort of static electricity, a little bit like when you’re rubbing a balloon across clothing.
But what we don’t really know is exactly why this is useful to the bees, if there is any reason why this is useful. Bees use static electricity when they’re picking up pollen. It helps pollen stick to their bodies. But that doesn’t really explain why a swarm should have such a high electric density.
KATHLEEN DAVIS: All right, so moving on from electric bees, let us talk about some really good news, which we don’t always get to talk about here in News Roundup. So after a really alarming spring and summer, cases of monkeypox are declining all over the place. Can you tell us a little bit about that?
TIM REVELL: Yeah, this is really great news. And there’s a really good article in Science this week by Kai Kupferschmidt on how all the signs for monkeypox are that the most recent outbreak in Europe and across the Americas is really on the decline. So in Europe, the World Health Organization reported about 2,000 cases a week in July. And now that’s down to just 100 cases a day. In the Americas, numbers have dropped by over half. And in the US in particular, it was about 400 cases a day in August. And now it’s just 30 a day.
KATHLEEN DAVIS: Wow, do we know why this drop is happening?
TIM REVELL: Yeah, so we can’t pinpoint exactly. But it’s three factors that are all related to each other. So one of them is the vaccines. And then we’ve also got behavior change and natural immunity. It seems that vaccines hasn’t played a huge role because quite a lot of this dip, this decrease, started happening before vaccines were rolled out in a major way. So the vaccines are only playing a minor role in at least the starting of the drop off.
What we think is happening is that probably this behavior change and natural immunity are the real big ones driving it. So a survey conducted by the CDC among men who have sex with men, which is the primary group monkeypox has been spreading in, that found that around half of people in that group had reduced their number of sexual contacts. And therefore, that just reduces the amount of opportunity monkeypox has to spread.
This is a really good example of the science working. When monkeypox was first on the rise in this most recent outbreak, there was lots of talk about how this is a controllable disease. With these three factors, vaccines, natural immunity, and behavior change, it could be brought under control. And we’ve seen that that is what’s happened.
KATHLEEN DAVIS: Meanwhile, Tim, daylight savings ends next weekend, which I am not looking forward to. But it has me thinking about clocks. And something very cool is happening in the world of clocks. It’s quantum. There are lasers involved. How does this work, Tim?
TIM REVELL: Yeah, so it’s not every week we get a completely new way to measure time. But that is the case this week. This is a really great story from my colleague Leah Crane at New Scientist. And this quantum watch, it measures time using quantum interference. It’s a phenomenon where, probably the most famous element of it, is shown in this thing called the double slit experiment, which shows that a single photon of light can pass through two slits at once. And then after going through those slits, the light interacts with itself in such a way that it causes this unique interference pattern on, say, a wall opposite.
And so rather than light, the quantum watch uses a cloud of helium atoms. And these, they interact with each other in such a way as to create an interference pattern, just like in the double slit experiment. But this changes in a very, very predictable way over time. And so by measuring this pattern, you can then work out how much time has elapsed since the interaction began. And it turns out this is an especially accurate way to measure time.
KATHLEEN DAVIS: OK, so I’m getting ready to put my Christmas list together, Tim. When can I get my very own laser helium quantum stopwatch?
TIM REVELL: I think the chances of this being delivered for Christmas this year are very slim indeed, unless you work in a physics lab. Where it is useful is in this particular type of experiment called a pump probe experiment, where you send a laser into a cloud of atoms. And so when you begin the experiment, the clock automatically begins, too. And so that means you have just an incredibly accurate way to measure how long the experiment has been going on for. And these experiments are really useful for designing things like new materials, and especially for solar panels.
KATHLEEN DAVIS: Oh, interesting. OK, we have one last story before we run out of time. In case anyone is watching too many scary movies and they’re getting nightmares, there’s a new tool for disarming recurring nightmares. Can you walk me through how this works?
TIM REVELL: Yeah, so recurring nightmares, if you experience these at the moment, one of the common ways to try and counteract that is through a process called imagery rehearsal therapy. Essentially, that involves imagining a more positive ending to the dream, and mentally rehearsing it during the day with a therapist, someone to help you through it. But that doesn’t always work for everyone.
And so in this new study, researchers found that one way to improve it is to add an associated sound to the positive experience. And then play it through a wireless headband whilst people were asleep. And that actually had a big effect in improving how this therapy works.
KATHLEEN DAVIS: So Tim, what kind of sound are we talking about here?
TIM REVELL: So it’s a very simple, neutral piano chord. And that was just played consistently during those positive ending imagination exercises. Then over a course of two weeks, they wore a headband during sleep, which then played the sound associated with the positive ending every 10 seconds during REM sleep, which is normally when you dream, for about two weeks.
And then the group in the headband group, they had a fifth as many nightmares as the control group. But the really interesting bit is that this continued after three months. After they had stopped wearing the headset, it seemed these effects were still in place.
KATHLEEN DAVIS: Tim Reevell is Deputy US Editor for New Scientist based in New York. Thank you so much.
TIM REVELL: Thanks for having me.