Telling Real from Fake in Disaster Times: A Noto Earthquake Case Study

Society Disaster

Social media provided fertile ground for disseminating misinformation in the wake of the New Year Noto earthquake. How can social media users discern truth from falsehood and avoid spreading groundless rumors themselves?

Patterns of Groundless Rumors During Disasters

Rumors spread online about the January 1, 2024, earthquake on the Noto Peninsula posed serious problems. The bulk were spread via X (Twitter), prompting Prime Minister Kishida Fumio to warn on January 2 that spreading malicious misinformation about the disaster would not be countenanced. Below I examine the various types of rumors that spread during a disaster, the patterns they follow, and tips for responding.

A huge disaster generates widespread unease in all of society, providing fertile ground for the spread of rumors. It was the same during the Great Kantō Earthquake of 1923, showing that well before the age of social media, humans have a propensity to engage in rumormongering in times of disaster. Today, the Internet and social media contribute exponentially to the speed and dissemination of rumors.

I believe that misinformation in times of disaster follows five patterns. I will describe these here, in the context of the Noto earthquake.

Rumors concerning the extent of damage

Scenes of the tsunami during the March 11, 2011, Great East Japan Earthquake presented as having occurred during the Noto disaster generated several million views. Unsubstantiated information about the causes of the earthquake, the fires that ravaged the centuries-old market in the city of Wajima, and the status of the Shika Nuclear Power Station operated by the Hokuriku Electric Power Company also circulated.

Rumors about criminal activity

Unfounded information spread that gangs of foreign thieves had flocked to Noto. There were also numerous rumors about looting and other criminal behavior at evacuation centers. As in the case of previous disasters, police authorities repeatedly called out this false information.

Fake rescue requests

Dubious rescue requests, such as “Help! I’m trapped in my car,” or “My friend’s family can’t get out of their house because the front door has warped,” spread rapidly without being verified. Hashtags like #SOS, #Pleaseretweet, or #Help make this type of message easy to repost, enabling circulation of false information about the true extent of the disaster.

Fake charity appeals

Some accounts requested donations via e-money, asking for funds for future reconstruction and other needs.

Conspiracy theories

Misinformation that the Noto Earthquake had been triggered artificially also spread. Analysis conducted by NHK revealed that in the roughly 24-hour period between the onset of the earthquake on January 1 and 5:30 pm on January 2, there had been 250,000 postings about an artificially triggered earthquake, some of which had garnered nearly 8.5 million views.

Rumors Have Profound Effects

Groundless rumors often contain elements appealing to the emotions or the desire to tell others. Strong emotions like anxiety or anger, or altruistic motives, may lead some individuals to disseminate rumors. A study by my research group has shown that many people engaging in such behavior do so because they feel anxious or believe that spreading a rumor is for the good of others or the community.

But rumormongering in times of disaster has profound effects. Not only can misinformation cause panic or confusion, it can also lead to confusion in evacuations, with some who should flee staying put and others who are safe staying in place taking to the roads, or improper handling of relief supplies. This can greatly complicate the situation in the affected areas and also negatively impact rescue efforts. Rescue teams may be deployed unnecessarily, placing a much heavier burden on local government employees.

Social divisions widen the gap between those who believe rumors and those who do not. This can exacerbate exclusion of or discrimination against certain groups, placing them at risk of vigilantism.

But the most negative effect of rumors is that they complicate obtaining accurate information. During a disaster, the responding parties base their decisions on the assumption that they are working with accurate information. Correct information encourages behavior to protect human life and also gives peace of mind to those affected by the disaster. But the spread of rumors lessens information reliability, making people skeptical and less able to make appropriate judgments.

The Pageview Hunt as a Catalyst

Some have pointed to the attention economy as a reason for rumors about the earthquake. In the modern digital space, the extent to which attention or interest are captured takes on economic value. In other words, when online information grows exponentially, overwhelming the human brain’s processing capacity, attracting people’s attention becomes more important than providing truthful information.

Winning pageviews as the be-all and end-all of online advertising is a good illustration of this concept. For websites supported by advertising revenue, pageviews are essential for bringing in as much advertising revenue as possible. As a result, those sites concentrate on attracting attention, not on the quality of the information they provide.

The attention economy has grown rapidly in recent years. It used to be that the traditional mass media and online media competed for eyeballs, but now, when everyone on the planet can blast out information, hits on websites set up by private individuals and YouTube views have become an income-generating source. The more outrageous the content, the easier it is to garner pageviews. This induces YouTubers and others to post more and more explosive material like exposés, dumb stunts, or citizen’s arrests in order to attract attention.

In August 2023, X (Twitter) inaugurated an advertising revenue-sharing program for platform users meeting certain criteria, for example, having 500 or more followers, or generating over 5 million impressions over a previous three-month period.

Hurdles to posting on X are much lower than for YouTube, and X can diffuse content much more broadly. This in turn makes it possible for even more people to participate in the attention economy. Consequently, X has become a massive conduit for disseminating all sorts of rumors drawing people’s attention. It has allowed the attention economy to penetrate among individuals, motivating them to post provocative material or disseminate misinformation, all for the sake of attracting more followers or gaining more impressions.

In the Noto earthquake, the numerous rumors posted online were copied and reposted, creating a far-spreading ripple of copied tweets. A considerable number of the reposting accounts were not ones that ordinarily posted in Japanese, the language of the reposted material. It is clear that posters, both domestic and foreign, were jumping in to earn impressions. Every time some cataclysm occurs, rumors are created or copied and widely spread.

AI Accelerates Social Turmoil

Generative AI has made it possible for any user to easily create fake images or videos and deep fakes. False or erroneous information can spread rapidly in this way: we are now entering a new “fakes 2.0” era. To take just one example, in the Israel-Hamas conflict, fake images or videos generated by AI are frequently posted in an effort to sway international opinion.

The spread of rumors is already affecting information-sharing in times of disaster. Drone photos of flooding in Shizuoka Prefecture in 2022 spread through social media, but the photos were in fact fakes generated by AI. Creating those photos required no particular technical expertise; it turned out that they had been posted by an individual using a service called Stable Diffusion, accessible to everyone. He later related that he had posted the images simply by fiddling with his smartphone from the comfort of his home.

Nowadays, pre-existing videos or images are often used to spread rumors, but I foresee that in the future, AI-generated false images or videos virtually indistinguishable from the genuine article will become prevalent. Advances in AI technology could greatly increase the volume of rumors in circulation and worsen social disorder.

Dealing with Rumors

To deal with rumors, we need to reflect on our own behavior, starting with the realization that we ourselves are susceptible to believing them. Thus, we must always be circumspect about the information we are exposed to.

In a study I conducted with my research group, 77.5% of respondents exposed to unverified rumors did not realize they had been hoodwinked. This was especially so among those in their fifties and sixties, compared with younger participants, demonstrating that rumors affect everyone, not just young people. We need to be aware that any of us could fall into the misinformation trap.

Research in the United States also found that most people were overconfident about their ability to tell truth from falsehood, and that they were the type of individual most likely to be misled.

It is also important to verify information. We should check its reliability by looking at how other media or individuals are treating the information or by conducting an image search. There are many ways of verifying sources, and while it is true that checking each one every single time is difficult when we are drowning in a flood of information, we need to pause and think for a moment to check before sharing information.

X/Twitter includes a “community notes” feature allowing users to flag misleading content and add more information. This feature lets contributors leave notes on any post; if enough contributors rate that note as helpful, the note will be publicly appended to a post. It may be helpful to check X’s community notes for substantiating information.

But rumors are not a problem just in the online sphere. My research has found that rumors spread considerably through direct conversations with family, friends or acquaintances. Communication research has also demonstrated that people find it easier to believe information that comes from those close to them rather than anything experts may say. We need to be alert to the risk of believing what people we know say without checking whether it is true. It is also likely that rumors spread via social media during disasters are further amplified through direct conversation.

Social media platform operators must do their best to ensure that reliable information is posted. The content-checking function of X has been weakened since the platform was acquired by Elon Musk. Respecting free speech is a given, but more stringent standards should apply to false information fueling social turmoil. One important countermeasure would be to curtail monetizing the dispersal of false information.

(Originally published in Japanese on March 14, 2024. Banner image © Pixta.)

disaster social media information technology