March 29, 2024

MIT Scientists Find Clues to Why Fake News Snowballs on Social Media

To shed some light on this murky subject, researchers at MIT established a theoretical design of a Twitter-like social network to study how news is shared and check out situations where a non-credible news item will spread out more widely than the reality.” We show that, even if people are rational in how they decide to share the news, this might still lead to the amplification of details with low trustworthiness.” They will make a cost-benefit analysis to see if, on average, this piece of news will move individuals closer to what they believe or move them away. Everyone has this cost, so the more extreme and the more fascinating the news is, the more you desire to share it,” Jadbabaie says.
Empirical research study by Sinan Aral at his collaborators at MIT reveals that false news is passed on more extensively than true news,” says Sanjeev Goyal, professor of economics at Cambridge University, who was not involved with this research study.

The researchers found that in such a setting, when a network is highly connected or the views of its members are dramatically polarized, news that is likely to be incorrect will spread out more widely and take a trip much deeper into the network than news with higher trustworthiness.
This theoretical work could notify empirical research studies of the relationship between news credibility and the size of its spread, which might assist social media business adapt networks to limit the spread of incorrect details.
” We show that, even if people are reasonable in how they choose to share the news, this could still lead to the amplification of information with low trustworthiness.,” states senior author Ali Jadbabaie, teacher and head of the Department of Civil and Environmental Engineering and a core faculty member of the Institute for Data, Systems, and Society (IDSS) and a primary investigator in the Laboratory for Information and Decision Systems (LIDS).
Joining Jadbabaie on the paper are very first author Chin-Chia Hsu, a graduate trainee in the Social and Engineering Systems program in IDSS, and Amir Ajorlou, a LIDS research scientist. The research study existed last week at the IEEE Conference on Decision and Control.
Contemplating persuasion
This research draws on a 2018 study by Sinan Aral, the David Austin Professor of Management at the MIT Sloan School of Management; Deb Roy, a teacher of media arts and sciences at the Media Lab; and previous postdoc Soroush Vosoughi (now an assistant teacher of computer technology at Dartmouth University). Their empirical study of data from Twitter found that incorrect news spreads broader, much faster, and much deeper than real news.
Jadbabaie and his partners wished to drill down on why this takes place.
They hypothesized that persuasion may be a strong motive for sharing news– possibly agents in the network wish to convince others to take on their perspective– and decided to build a theoretical design that would let them explore this possibility.
In their design, representatives have some prior belief about a policy, and their objective is to convince fans to move their beliefs closer to the agents side of the spectrum.
A news item is initially launched to a little, random subgroup of representatives, which need to decide whether to share this news with their fans. An agent weighs the newsworthiness of the product and its reliability, and updates its belief based upon how unexpected or persuading the news is.
” They will make a cost-benefit analysis to see if, on average, this piece of news will move people closer to what they think or move them away. Or a reputation expense might come if I share something that is humiliating. Everyone has this cost, so the more extreme and the more fascinating the news is, the more you desire to share it,” Jadbabaie says.
The agent will constantly share the news if the news verifies the representatives perspective and has persuasive power that outweighs the small cost. If an agent believes the news product is something others might have already seen, the agent is disincentivized to share it.
Since a representatives desire to share news is a product of its viewpoint and how persuasive the news is, the more severe an agents point of view or the more surprising the news, the most likely the representative will share it.
The scientists used this model to study how details spreads during a news waterfall, which is an unbroken sharing chain that rapidly permeates the network.
Connectivity and polarization
The group discovered that when a network has high connectivity and the news is unexpected, the trustworthiness threshold for beginning a news waterfall is lower. High connectivity suggests that there are numerous connections between numerous users in the network.
Similarly, when the network is largely polarized, there are lots of agents with extreme views who wish to share the news product, starting a news waterfall. In both these instances, news with low trustworthiness develops the biggest cascades.
” For any piece of news, there is a natural network speed limit, a series of connection, that assists in good transmission of details where the size of the cascade is maximized by true news. If you surpass that speed limit, you will get into circumstances where unreliable news or news with low reliability has a bigger cascade size,” Jadbabaie states.
It is less likely that a poorly reliable piece of news will spread out more commonly than the fact if the views of users in the network become more diverse.
Jadbabaie and his colleagues designed the representatives in the network to act reasonably, so the design would better catch actions real humans might take if they desire to persuade others.
” Someone might say that is not why individuals share, and that is legitimate. Why people do particular things is a subject of intense debate in cognitive science, social psychology, neuroscience, economics, and political science,” he says.
Their design also demonstrates how expenses can be controlled to lower the spread of incorrect info. If the expense to do so outweighs the advantage of sharing, representatives make a cost-benefit analysis and wont share news.
” We do not make any policy prescriptions, but one thing this work recommends is that, perhaps, having actually some expense related to sharing news is not a bad idea. The factor you get lots of these cascades is because the expense of sharing the news is actually extremely low,” he says.
” The function of socials media in affecting and forming opinions behavior has actually been widely noted. Empirical research study by Sinan Aral at his collaborators at MIT shows that incorrect news is handed down more widely than real news,” says Sanjeev Goyal, teacher of economics at Cambridge University, who was not involved with this research study. “In their brand-new paper, Ali Jadbabaie and his partners provide us an explanation for this puzzle with the aid of an elegant model.”
Recommendation: “Persuasion, News Sharing, and Cascades on Social Networks” by Chin-Chia Hsu, Amir Ajorlou and Ali Jadbabaie, 1 October 2021, SSRN.DOI: 10.2139/ ssrn.3934010.
This work was supported by an Army Research Office Multidisciplinary University Research Initiative grant and a Vannevar Bush Fellowship from the Office of the Secretary of Defense.

MIT scientists constructed a theoretical design to study how news spreads on a Twitter-like social media and found that when a network is highly connected or when the views of its members are sharply polarized, false news will spread wider than news that is seen as more credible. Credit: Jose-Luis Olivares, MIT
A new design shows that the more polarized and hyperconnected a social media is, the more most likely misinformation will spread out.
The spread of false information on social networks is a pressing societal issue that tech companies and policymakers continue to come to grips with, yet those who study this concern still do not have a deep understanding of why and how false news spreads.
To shed some light on this dirty subject, scientists at MIT established a theoretical model of a Twitter-like social media network to study how news is shared and explore situations where a non-credible news product will spread more commonly than the truth. Agents in the design are driven by a desire to encourage others to handle their perspective: The essential assumption in the model is that people bother to share something with their fans if they think it is persuasive and most likely to move others closer to their state of mind. Otherwise they wont share.