Machine learning algorithms significantly outperform human judgment in detecting lying during high-stakes strategic interactions, according to new research from the University of California San Diego’s Rady School of Management.

The study can have major implications for the spread of misinformation, as machine learning could be used to bolster efforts to reduce fictitious content on major platforms like YouTube, Tik-Tok and Instagram.

The study, to be published in Management Science and available as a working paper, focused on participants’ ability to detect lying on the popular British TV show “Golden Balls,” which aired from 2007 to 2010. It finds that while humans struggle to predict contestants’ deception behavior, algorithms perform much better.

“We find that there are certain ‘tells’ when a person is being deceptive,” said Marta Serra-Garcia, lead author of the study and associate professor of behavioral economics at the UC San Diego Rady School of Management.

“For example, if someone is happier, they are telling the truth and there are other visual, verbal, vocal cues that we as humans all share when we are being honest and telling the truth. Algorithms work better at uncovering these correlations.”

The algorithms used in the research achieved an impressive accuracy rate, correctly predicting contestant behavior 74% of the time, compared to the 51%–53% accuracy rate achieved by the more than 600 humans who participated in the study.

In addition to comparing machine learning and human abilities to detect deception, the study also tested how algorithms could be leveraged to help people better tell apart those who lie and those who tell the truth.

In one experiment, two different groups of study participants watched the same set of “Golden Balls” episodes. One group had the videos flagged by machine learning before they viewed them. The flags indicated that the algorithm predicted the contestant was most likely lying.

Another group watched the same video and after they viewed it, they were told the algorithm flagged the video for deception. Participants were much more likely to trust the machine learnings’ insights and better predict lying, if they got the flag message before watching the video.

“Timing is crucial when it comes to the adoption of algorithmic advice,” said Serra-Garcia. “Our findings show that participants are far more likely to rely on algorithmic insights when these are presented early in the decision-making process. This has particular importance for online platforms like YouTube and TikTok, which can use algorithms to flag potentially deceptive content.”

Co-author Uri Gneezy, professor of behavioral economics at the Rady School added, “Our study suggests that these online platforms could improve the effectiveness of their flagging systems by presenting algorithmic warnings before users engage with the content, rather than after, which could lead to misinformation spreading less rapidly.”

Some of these social media websites are already using algorithms to detect suspicious content, but in many cases, a video has to be reported by a user and then investigated by staff who can flag the content or take it down. These processes can be drawn out, as employees at tech companies like TikTok get overburdened with investigations.

The authors conclude, “Our study shows how technology can enhance human decision making and it’s an example of how humans can interact with AI when AI can be helpful. We hope the findings can help organizations and platforms better design and deploy machine learning tools, especially in situations where accurate decision-making is critical.”

Volunteering used to be a mainstay of U.S. culture. But in recent years, giving back to their community hasn’t played as big a role in many Americans’ lives.

New research from the University of Georgia suggests the economy may be to blame.

The researchers found that people living in disadvantaged communities or areas that have high levels of economic inequality were less likely to volunteer.

“Historically, rural areas have had higher volunteering rates than urban ones,” said Rebecca Nesbit, lead author of the study and a professor in UGA’s School of Public and International Affairs.

“These communities often have closer ties and more social interaction with each other, and those close ties may make them more likely to volunteer. Because when you’re volunteering for the local food bank in these communities, you’re helping people that you have a personal connection to.”

The recession of 2008 didn’t help matters. And more than a decade and a half later, volunteering rates have yet to recover.

“Any advantage to volunteering afforded by good economic growth before the recession was wiped out after the recession, and that can lead people to change their behavior,” Nesbit said. “In poor economic conditions, people might take energy away from their voluntary activities to put it into more income-producing activities that create a greater sense of personal stability.”

Economically disadvantaged areas hit the hardest

The study is the first analysis of confidential level volunteering data in a secure U.S. Census Bureau Research Data Center, which is a nationally representative sample of 56,000 households interviewed each month. This data is considered the premiere source of information on national and state volunteering statistics.

The researchers relied on a dataset of about 90,000 individuals for each year of the survey.

The study examined effects of economic disadvantage and inequality, and how the Great Recession exacerbated existing differences in volunteering rates between rural and urban communities.

The researchers found that the recession had the biggest dampening effect on volunteering in areas with the most economic growth and above average income equality.

“What the decline in volunteering tells us is that there are a lot of communities that were hit by the recession and just haven’t bounced back,” Nesbit said. “Whether that decline is going to continue or whether these communities will rebound eventually, we don’t know yet.”

The demographics of rural communities are changing, with more youth leaving their hometowns for bigger cities while the population left behind ages. That shift in people’s sense of community may be one reason these communities’ volunteering rates dropped so significantly, the researchers said.

People who lived in areas with growing economies were more likely to volunteer, according to the study.

Even years after the recession ended, the negative effects persisted, which may indicate lingering social or psychological effects that make people more hesitant to invest time and resources into volunteer efforts, the researchers said.

“One general high-level finding is that local economic conditions matter for volunteering. That’s something we can’t ignore,” Nesbit said. “An implication of that is that as we talk about economic development for communities, we shouldn’t divorce that from the civic development of communities.

“Policymakers need to understand that if we want to strengthen communities, particularly these rural communities, we need a more holistic approach. It can’t just be about economic development, and it can’t just be about civic engagement. It has to be both.”

The increasing popularity of cryptocurrencies has turned social media into a central place where users look for help when they have problems with their crypto wallet or private key. Scammers take advantage of this situation to make money with fake support offerings or to gain access to wallets or keys.

CISPA researcher Dr. Bhupendra Acharya has presented the first large-scale study on how these scams work and provided an end-to-end analysis of the scam operations in X (formerly known as Twitter). He presented his findings at the 45th IEEE Symposium on Security and Privacy in May 2024. The paper is available on the arXiv preprint server.

Cryptocurrencies such as Bitcoin or Ethereum are widely gaining acceptance because of their decentralized nature and because they grant anonymity to their users. In order to manage and sell cryptocurrencies, users need so-called crypto wallets, which basically are digital wallets for cryptocurrencies.

The best-known wallets are Metamask, Coinbase and Trust. In order to access these wallets, secret keys are required. Anyone with access to the secret keys can manage or access the crypto wallets. In the event of secret key loss, the crypto wallets remain inaccessible.

“We noticed that, as cryptocurrencies have become more popular, people have also been talking about them on social media. This also includes technical support issues such as wallet inaccess, loss of private key phrases, etc., which attracts fraudsters who fake technical support, effectively impersonating official support,” explains Acharya.

Many people prefer to seek help in a chat group or via a tweet instead of contacting the official support channels of the respective crypto wallet provider directly. “In our study, we uncovered how scammers exploit users in social media to either gain access to crypto wallets or simply ask for payment in return for technical support they are faking,” says Acharya.

On the scammers’ trail with HoneyTweet

In order to investigate how support scams in social media actually work, Acharya developed a tool called HoneyTweet. “HoneyTweet automatically sends out unique tweets with keywords for technical support requests in order to bait scammers,” Acharya explains.

“Scammers offering fake support are contacted via a semi-automated tool to detect the scamming payment methods or the modus operandi of scammers,” he continues. The scammers come up with various fake offers such as the software tool “Zeus,” which they claim will retrieve wallet access, and ask for money as part of the support.

Users were often pivoted to external communication channels during the conversation to avoid scam detection on the original platform. With the aid of HoneyTweet, Acharya and his colleagues baited more than 9,000 scammers within three months and traced them on six social media platforms including PayPal and cryptocurrency addresses, which were used as scamming payment methods.

The most important results of the study

In their study, Acharya and his colleagues were able to show that support scams for crypto wallets are a widespread phenomenon on social media such as X. “We found that social media still has some work to do in order to stop these scams,” Acharya says. “And we also found out that scammers often use several social media platforms for their scam attempts.

“Beyond X, the scammers ask to be contacted via direct messages on Instagram, Facebook, Telegram, WhatsApp and others.”

Basically, the scammers work in chain operations that link several social media platforms. During the scam process, the scammers first try to build trust and later perform social engineering tricks, initiating direct message communication where the actual scams take place.

Upon direct messaging, the potential victim is asked to either release their private key or pay for the “fake” support via the scammer’s provided payment method. By collaborating with PayPal and sharing the detected scam accounts with the payment service provider, the researchers were able to further validate the scam’s financial impact.

Takeaways for businesses and users

“There are two groups that could adopt our recommendations,” Acharya explains. “The first one consists of the involved services, like the crypto wallet providers. They should monitor all activity directly associated with their brand name and take action if scammers attempt to impersonate their brand. The second group consists of social media like X, Instagram, Facebook, Telegram and others.

“It is important to jointly monitor what is going on in terms of scam chains, because the scam does not necessarily occur on the platform where the chat started out. The final scam might take place at end of the chain, i.e., on another platform. In order to combat those chains, cooperation between the social media services is particularly important.”

Additionally, users of crypto wallets can also take action. Acharya recommends making sure to engage only with official providers of cryptocurrency wallets and be cautious with all unofficial channels. In no case should the information be shared via Google Forms or similar platforms.

“Crypto wallets or social media accounts affiliated with official crypto wallets will never ask their users for their secret keys,” the CISPA researcher concludes.

The future belongs to (secure) digital currencies

Acharya, who during the conversation reveals himself to be a big fan of digital currencies and a cryptocurrency user, sees a lot of potential in cryptocurrencies. “I believe that digital currencies like cryptocurrencies are the next generation of currencies and that they will replace existing currencies in the future,” he is convinced. “However, what we need is a system that is secure enough to create and operate a digital currency.”

As a researcher, he wants to continue contributing to this goal. “One project is using ChatGPT to chat with the scammers based on HoneyTweet,” he explains. “In this context, we also focus on different categories of fraud, such as alleged account recovery.

“In another follow-up study, we will use a deepfake-based method to chat and communicate with the scammers via Zoom video and phone with the aim of identifying further types of fraud mechanisms.”

It will be exciting to see what fraud mechanisms in the area of cryptocurrencies will be uncovered by Acharya and his colleagues.

Provided by
CISPA Helmholtz Center for Information Security

Researchers have just published their findings on the analysis of economic activities over a century of urban growth in the city of Paris.

The study is published in Nature Cities. The authors are Julie Gravier (Center de Recherches Historiques and Center d’Analyse et de Mathématique Sociales, EHESS) and Marc Barthelemy (Institut de Physique Théorique and Center d’Analyse et de Mathématique Sociales, EHESS).

Understanding the evolution of economic activities alongside the growth of large cities is crucial but challenging, particularly for historical periods where comprehensive data is often unavailable. In this study, the authors leverage a newly released dataset, constructed from city directories, which includes extensive records of individuals, merchants, businesses, organizations, and institutions in Paris from 1829 to 1907.

This dataset contains approximately 1 million entries of economic activities, regularly updated over a span of 79 years, offering a rare historical perspective on urban growth and economic evolution. This period also encompasses significant disruptions, such as the transformative Haussmann’s works (1853–1870) and political upheavals like the Franco-Prussian War and the Commune revolution (1870–1871).

Contemporary research on the dynamics of economic activity in cities typically spans shorter timeframes—ranging from a few years to a few decades. In contrast, the analysis of Barthelemy and Gravier covers nearly a century, providing a deeper understanding of how economic activities evolved during Paris’ expansion.

The authors showed in the case of Paris during the 19th century that not all activities are equivalent and that they can be grouped in categories with different dynamics and scaling. In particular, it seems that some activities naturally accompany the growth of the city and can be considered as intrinsic to its development.

These activities essentially answer to the basic needs of the residents, and the type of scaling with population of the number of these activities depends essentially on their underlying logic: It seems intuitive that the number of food stores scales more or less linearly with population while the number of administrative or educational needs follow an optimization logic (and scale sublinearly).

All these activities constitute the core of the city, while other activities determine its specific features. These quickly developing activities usually respond to some passing fad or to a need of the specific period and the phase of the city’s development.

These findings demonstrate that economic activities scale differently with population and suggest that while some activities are fundamental to a city’s growth, others become prominent during periods of innovation or specialization.

Additionally, Paris’ evolution should be viewed within the broader context of an urban system, as the largest city in France at the time. Urban scaling laws, which help interpret the relationship between city size and economic activity, suggest that activities with scaling exponents greater than 1 represent early stages of innovation adoption, often concentrated in large cities. Conversely, activities with exponents below 1 correspond to more mature, established activities.

The typology presented in the paper likely extends to other growing cities, highlighting how different types of activities contribute uniquely to urban development, depending on the city’s phase of growth or specialization. This analysis provides new insights into the diverse roles of economic activities in shaping the evolution of large cities.