This article first appeared in The Checkup, MIT Technology Review’s weekly biotech newsletter. To receive it in your inbox every Thursday, and read articles like this first, sign up here. 

Can you spot a liar? It’s a question I imagine has been on a lot of minds lately, in the wake of various televised political debates. Research has shown that we’re generally pretty bad at telling a truth from a lie.
 
Some believe that AI could help improve our odds, and do better than dodgy old fashioned techniques like polygraph tests. AI-based lie detection systems could one day be used to help us sift fact from fake news, evaluate claims, and potentially even spot fibs and exaggerations in job applications. The question is whether we will trust them. And if we should.

In a recent study, Alicia von Schenk and her colleagues developed a tool that was significantly better than people at spotting lies. Von Schenk, an economist at the University of Würzburg in Germany, and her team then ran some experiments to find out how people used it. In some ways, the tool was helpful—the people who made use of it were better at spotting lies. But they also led people to make a lot more accusations.

In their study published in the journal iScience, von Schenk and her colleagues asked volunteers to write statements about their weekend plans. Half the time, people were incentivized to lie; a believable yet untrue statement was rewarded with a small financial payout. In total, the team collected 1,536 statements from 768 people.
 
They then used 80% of these statements to train an algorithm on lies and truths, using Google’s AI language model BERT. When they tested the resulting tool on the final 20% of statements, they found it could successfully tell whether a statement was true or false 67% of the time. That’s significantly better than a typical human; we usually only get it right around half the time.
 
To find out how people might make use of AI to help them spot lies, von Schenk and her colleagues split 2,040 other volunteers into smaller groups and ran a series of tests.
 
One test revealed that when people are given the option to pay a small fee to use an AI tool that can help them detect lies—and earn financial rewards—they still aren’t all that keen on using it. Only a third of the volunteers given that option decided to use the AI tool, possibly because they’re skeptical of the technology, says von Schenk. (They might also be overly optimistic about their own lie-detection skills, she adds.)
 
But that one-third of people really put their trust in the technology. “When you make the active choice to rely on the technology, we see that people almost always follow the prediction of the AI… they rely very much on its predictions,” says von Schenk.

This reliance can shape our behavior. Normally, people tend to assume others are telling the truth. That was borne out in this study—even though the volunteers knew half of the statements were lies, they only marked out 19% of them as such. But that changed when people chose to make use of the AI tool: the accusation rate rose to 58%.
 
In some ways, this is a good thing—these tools can help us spot more of the lies we come across in our lives, like the misinformation we might come across on social media.
 
But it’s not all good. It could also undermine trust, a fundamental aspect of human behavior that helps us form relationships. If the price of accurate judgements is the deterioration of social bonds, is it worth it?
 
And then there’s the question of accuracy. In their study, von Schenk and her colleagues were only interested in creating a tool that was better than humans at lie detection. That isn’t too difficult, given how terrible we are at it. But she also imagines a tool like hers being used to routinely assess the truthfulness of social media posts, or hunt for fake details in a job hunter’s resume or interview responses. In cases like these, it’s not enough for a technology to just be “better than human” if it’s going to be making more accusations. 
 
Would we be willing to accept an accuracy rate of 80%, where only four out of every five assessed statements would be correctly interpreted as true or false? Would even 99% accuracy suffice? I’m not sure.
 
It’s worth remembering the fallibility of historical lie detection techniques. The polygraph was designed to measure heart rate and other signs of “arousal” because it was thought some signs of stress were unique to liars. They’re not. And we’ve known that for a long time. That’s why lie detector results are generally not admissible in US court cases. Despite that, polygraph lie detector tests have endured in some settings, and have caused plenty of harm when they’ve been used to hurl accusations at people who fail them on reality TV shows.
 
Imperfect AI tools stand to have an even greater impact because they are so easy to scale, says von Schenk. You can only polygraph so many people in a day. The scope for AI lie detection is almost limitless by comparison.
 
“Given that we have so much fake news and disinformation spreading, there is a benefit to these technologies,” says von Schenk. “However, you really need to test them—you need to make sure they are substantially better than humans.” If an AI lie detector is generating a lot of accusations, we might be better off not using it at all, she says.

---

Now read the rest of The Checkup

Read more from MIT Technology Review’s archive

AI lie detectors have also been developed to look for facial patterns of movement and “microgestures” associated with deception. As Jake Bittle puts it: “the dream of a perfect lie detector just won’t die, especially when glossed over with the sheen of AI.”
 
On the other hand, AI is also being used to generate plenty of disinformation. As of October last year, generative AI was already being used in at least 16 countries to “sow doubt, smear opponents, or influence public debate,” as Tate Ryan-Mosley reported.
 
The way AI language models are developed can heavily influence the way that they work. As a result, these models have picked up different political biases, as my colleague Melissa Heikkilä covered last year.
 
AI, like social media, has the potential for good or ill. In both cases, the regulatory limits we place on these technologies will determine which way the sword falls, argue Nathan E. Sanders and Bruce Schneier.
 
Chatbot answers are all made up. But there’s a tool that can give a reliability score to large language model outputs, helping users work out how trustworthy they are. Or, as Will Douglas Heaven put it in an article published a few months ago, a BS-o-meter for chatbots.

From around the web

Scientists, ethicists and legal experts in the UK have published a new set of guidelines for research on synthetic embryos, or, as they call them, “stem cell-based embryo models (SCBEMs).” There should be limits on how long they are grown in labs, and they should not be transferred into the uterus of a human or animal, the guideline states. They also note that, if, in future, these structures look like they might have the potential to develop into a fetus, we should stop calling them “models” and instead refer to them as “embryos.”

Antimicrobial resistance is already responsible for 700,000 deaths every year, and could claim 10 million lives per year by 2050. Overuse of broad spectrum antibiotics is partly to blame. Is it time to tax these drugs to limit demand? (International Journal of Industrial Organization)

Spaceflight can alter the human brain, reorganizing gray and white matter and causing the brain to shift upwards in the skull. We need to better understand these effects, and the impact of cosmic radiation on our brains, before we send people to Mars. (The Lancet Neurology)

The vagus nerve has become an unlikely star of social media, thanks to influencers who drum up the benefits of stimulating it. Unfortunately, the science doesn’t stack up. (New Scientist)

A hospital in Texas is set to become the first in the country to enable doctors to see their patients via hologram. Crescent Regional Hospital in Lancaster has installed Holobox—a system that projects a life-sized hologram of a doctor for patient consultations. (ABC News)

This article first appeared in The Checkup, MIT Technology Review’s weekly biotech newsletter. To receive it in your inbox every Thursday, and read articles like this first, sign up here. 

This week I’ve been thinking about thought. It was all brought on by reading my colleague Niall Firth’s recent cover story about the use of artificial intelligence in video games. The piece describes how game companies are working to incorporate AI into their products to create more immersive experiences for players.

These companies are applying large language models to generate new game characters with detailed backstories—characters that could engage with a player in any number of ways. Enter in a few personality traits, catchphrases, and other details, and you can create a background character capable of endless unscripted, never-repeating conversations with you.

This is what got me thinking. Neuroscientists and psychologists have long been using games as research tools to learn about the human mind. Numerous video games have been either co-opted or especially designed to study how people learn, navigate, and cooperate with others, for example. Might AI video games allow us to probe more deeply, and unravel enduring mysteries about our brains and behavior?

I decided to call up Hugo Spiers to find out. Spiers is a neuroscientist at University College London who has been using a game to study how people find their way around. In 2016, Spiers and his colleagues worked with Deutsche Telekom and the games company Glitchers to develop Sea Hero Quest, a mobile video game in which players have to navigate a sea in a boat. They have since been using the game to learn more about how people lose navigational skills in the early stages of Alzheimer’s disease.

The use of video games in neuroscientific research kicked into gear in the 1990s, Spiers tells me, following the release of 3D games like Wolfenstein 3D and Duke Nukem. “For the first time, you could have an entirely simulated world in which to test people,” he says.

Scientists could observe and study how players behaved in these games: how they explored their virtual environment, how they sought rewards, how they made decisions. And research volunteers didn’t need to travel to a lab—their gaming behavior could be observed from wherever they happened to be playing, whether that was at home, at a library, or even inside an MRI scanner.

For scientists like Spiers, one of the biggest advantages of using games in research is that people want to play them. The use of games allows scientists to explore fundamental experiences like fun and curiosity. Researchers often offer a small financial incentive to volunteers who take part in their studies. But they don’t have to pay people to play games, says Spiers.

You’re much more likely to have fun if you’re motivated. It’s just not quite the same when you’re doing something purely for the money. And not having to pay participants allows researchers to perform huge studies on smaller budgets. Spiers has been able to collect data on over 4 million people from 195 countries, all of whom have willingly played Sea Hero Quest.  

AI could help researchers go even further. A rich, immersive world filled with characters that interact in realistic ways could help them study how our minds respond to various social settings and how we relate to other individuals. By observing how players interact with AI characters, scientists can learn more about how we cooperate—and compete—with others. It would be far cheaper and easier than hiring actors to engage with research volunteers, says Spiers.

Spiers himself is interested in learning how people hunt, whether for food, clothes, or a missing pet. “We still use these bits of our brain that our ancestors would have used daily, and of course some traditional communities still hunt,” he tells me. “But we know almost nothing about how the brain does this.” He envisions using AI-driven nonplayer characters to learn more about how humans cooperate for hunting.

There are other, newer questions to explore. At a time when people are growing attached to “virtual companions,” and an increasing number of AI girlfriends and boyfriends are being made available, AI video-game characters could also help us understand these novel relationships. “People are forming a relationship with an artificial agent,” says Spiers. “That’s inherently interesting. Why would you not want to study that?”

---

Now read the rest of The Checkup

Read more from MIT Technology Review’s archive:

My fellow London-based colleagues had a lot of fun generating an AI game character based on Niall. He turned out to be a sarcastic, smug, and sassy monster.

Google DeepMind has developed a generative AI model that can generate a basic but playable video game from a short description, a hand-drawn sketch, or a photo, as my colleague Will Heaven wrote earlier this year. The resulting games look a bit like Super Mario Bros.

Today’s world is undeniably gamified, argues Bryan Gardiner. He explores how we got here in another article from the Play issue of the magazine.

Large language models behave in unexpected ways. And no one really knows why, as Will wrote in March.

Technologies can be used to study the brain in lots of different ways—some of which are much more invasive than others. Tech that aims to read your mind and probe your memories is already being used, as I wrote in a previous edition of The Checkup.

From around the web:

Bad night of sleep left you needing a pick-me-up? Scientists have designed an algorithm to deliver tailored sleep-and-caffeine-dosing schedules to help tired individuals “maximize the benefits of limited sleep opportunities and consume the least required amount of caffeine.” (Yes, it may have been developed with the US Army in mind, but surely we all stand to benefit?) (Sleep)

Is dog cloning a sweet way to honor the memory of a dearly departed pet, or a “frivolous and wasteful and ethically obnoxious” pursuit in which humans treat living creatures as nothing more than their own “stuff”? This feature left me leaning toward the latter view, especially after learning that people tend to like having dogs with health problems … (The New Yorker)

States that have enacted the strongest restrictions to abortion access have also seen prescriptions for oral contraceptives plummet, according to new research. (Mother Jones)

And another study has linked Texas’s 2021 ban on abortion in early pregnancy with an increase in the number of infant deaths recorded in the state. In 2022, across the rest of the US, the number of infant deaths ascribed to anomalies present at birth decreased by 3.1%. In Texas, this figure increased by 22.9%. (JAMA Pediatrics)

We are three months into the bird flu outbreak in US dairy cattle. But the country still hasn’t implemented a sufficient testing infrastructure and doesn’t fully understand how the virus is spreading. (STAT)

This article first appeared in The Checkup, MIT Technology Review’s weekly biotech newsletter. To receive it in your inbox every Thursday, and read articles like this first, sign up here. 

Earlier this week, the US surgeon general, also known as the “nation’s doctor,” authored an article making the case that health warnings should accompany social media. The goal: to protect teenagers from its harmful effects. “Adolescents who spend more than three hours a day on social media face double the risk of anxiety and depression symptoms,” Vivek Murthy wrote in a piece published in the New York Times. “Additionally, nearly half of adolescents say social media makes them feel worse about their bodies.”

His concern instinctively resonates with me. I’m in my late 30s, and even I can end up feeling a lot worse about myself after a brief stint on Instagram. I have two young daughters, and I worry about how I’ll respond when they reach adolescence and start asking for access to whatever social media site their peers are using. My children already have a fascination with cell phones; the eldest, who is almost six, will often come into my bedroom at the crack of dawn, find my husband’s phone, and somehow figure out how to blast “Happy Xmas (War Is Over)” at full volume.

But I also know that the relationship between this technology and health isn’t black and white. Social media can affect users in different ways—often positively. So let’s take a closer look at the concerns, the evidence behind them, and how best to tackle them.

Murthy’s concerns aren’t new, of course. In fact, almost any time we are introduced to a new technology, some will warn of its potential dangers. Innovations like the printing press, radio, and television all had their critics back in the day. In 2009, the Daily Mail linked Facebook use to cancer.

More recently, concerns about social media have centered on young people. There’s a lot going on in our teenage years as our brains undergo maturation, our hormones shift, and we explore new ways to form relationships with others. We’re thought to be more vulnerable to mental-health disorders during this period too. Around half of such disorders are thought to develop by the age of 14, and suicide is the fourth-leading cause of death in people aged between 15 and 19, according to the World Health Organization. Many have claimed that social media only makes things worse.

Reports have variously cited cyberbullying, exposure to violent or harmful content, and the promotion of unrealistic body standards, for example, as potential key triggers of low mood and disorders like anxiety and depression. There have also been several high-profile cases of self-harm and suicide with links to social media use, often involving online bullying and abuse. Just this week, the suicide of an 18-year-old in Kerala, India, was linked to cyberbullying. And children have died after taking part in dangerous online challenges made viral on social media, whether from inhaling toxic substances, consuming ultra-spicy tortilla chips, or choking themselves.

Murthy’s new article follows an advisory on social media and youth mental health published by his office in 2023. The 25-page document, which lays out some of known benefits and harms of social media use as well as the “unknowns,” was intended to raise awareness of social media as a health issue. The problem is that things are not entirely clear cut.

“The evidence is currently quite limited,” says Ruth Plackett, a researcher at University College London who studies the impact of social media on mental health in young people. A lot of the research on social media and mental health is correlational. It doesn’t show that social media use causes mental health disorders, Plackett says.

The surgeon general’s advisory cites some of these correlational studies. It also points to survey-based studies, including one looking at mental well-being among college students after the rollout of Facebook in the mid-2000s. But even if you accept the authors’ conclusion that Facebook had a negative impact on the students’ mental health, it doesn’t mean that other social media platforms will have the same effect on other young people. Even Facebook, and the way we use it, has changed a lot in the last 20 years.

Other studies have found that social media has no effect on mental health. In a study published last year, Plackett and her colleagues surveyed 3,228 children in the UK to see how their social media use and mental well-being changed over time. The children were first surveyed when they were aged between 12 and 13, and again when they were 14 to 15 years old.

Plackett expected to find that social media use would harm the young participants. But when she conducted the second round of questionnaires, she found that was not the case. “Time spent on social media was not related to mental-health outcomes two years later,” she tells me.

Other research has found that social media use can be beneficial to young people, especially those from minority groups. It can help some avoid loneliness, strengthen relationships with their peers, and find a safe space to express their identities, says Plackett. Social media isn’t only for socializing, either. Today, young people use these platforms for news, entertainment, school, and even (in the case of influencers) business.

“It’s such a mixed bag of evidence,” says Plackett. “I’d say it’s hard to draw much of a conclusion at the minute.”

In his article, Murthy calls for a warning label to be applied to social media platforms, stating that “social media is associated with significant mental-health harms for adolescents.”

But while Murthy draws comparisons to the effectiveness of warning labels on tobacco products, bingeing on social media doesn’t have the same health risks as chain-smoking cigarettes. We have plenty of strong evidence linking smoking to a range of diseases, including gum disease, emphysema, and lung cancer, among others. We know that smoking can shorten a person’s life expectancy. We can’t make any such claims about social media, no matter what was written in that Daily Mail article.

Health warnings aren’t the only way to prevent any potential harms associated with social media use, as Murthy himself acknowledges. Tech companies could go further in reducing or eliminating violent and harmful content, for a start. And digital literacy education could help inform children and their caregivers how to alter the settings on various social media platforms to better control the content children see, and teach them how to assess the content that does make it to their screens.

I like the sound of these measures. They might even help me put an end to the early-morning Christmas songs. 

---

Now read the rest of The Checkup

Read more from MIT Technology Review’s archive:

Bills designed to make the internet safer for children have been popping up across the US. But individual states take different approaches, leaving the resulting picture a mess, as Tate Ryan-Mosley explored.

Dozens of US states sued Meta, the parent company of Facebook, last October. As Tate wrote at the time, the states claimed that the company knowingly harmed young users, misled them about safety features and harmful content, and violated laws on children’s privacy.  

China has been implementing increasingly tight controls over how children use the internet. In August last year, the country’s cyberspace administrator issued detailed guidelines that include, for example, a rule to limit use of smart devices to 40 minutes a day for children under the age of eight. And even that use should be limited to content about “elementary education, hobbies and interests, and liberal arts education.” My colleague Zeyi Yang had the story in a previous edition of his weekly newsletter, China Report.

Last year, TikTok set a 60-minute-per-day limit for users under the age of 18. But the Chinese domestic version of the app, Douyin, has even tighter controls, as Zeyi wrote last March.

One way that social media can benefit young people is by allowing them to express their identities in a safe space. Filters that superficially alter a person’s appearance to make it more feminine or masculine can help trans people play with gender expression, as Elizabeth Anne Brown wrote in 2022. She quoted Josie, a trans woman in her early 30s. “The Snapchat girl filter was the final straw in dropping a decade’s worth of repression,” Josie said. “[I] saw something that looked more ‘me’ than anything in a mirror, and I couldn’t go back.”

From around the web

Could gentle shock waves help regenerate heart tissue? A trial of what’s being dubbed a “space hairdryer” suggests the treatment could help people recover from bypass surgery. (BBC)

“We don’t know what’s going on with this virus coming out of China right now.” Anthony Fauci gives his insider account of the first three months of the covid-19 pandemic. (The Atlantic)

Microplastics are everywhere. It was only a matter of time before scientists found them in men’s penises. (The Guardian)

Is the singularity nearer? Ray Kurzweil believes so. He also thinks medical nanobots will allow us to live beyond 120. (Wired)