By Hiroyuki Fujishiro, Associate Professor at Hosei University and the Founder of the Japan Center of Education for Journalists
In Japan, the dissemination of mis- and disinformation has been accelerated by online media platforms including one of the most popular news portal sites in the country, while traditional media channels continue to play down its impact on social media users. During the country’s lower house election in late 2017, the Japan Center of Education for Journalists (JCEJ) and a dozen of students at Hosei University launched a collaborative verification project to delve into the ecosystem surrounding the so-called “fake news” in Japan.
The project was launched as soon as Prime Minister Shinzo Abe resolved the House of Representatives on Sept. 28, 2017. Through Oct. 23, 2017, a total of nine Hosei University students majoring in Media Studies monitored Twitter and Facebook every day in search for questionable information associated with the election. The students used various queries including names of political parties and candidates to search for possible misinformation.
A total of 275 items of social media content were collected over the course of roughly a month. Each student limited the time of web searching to one hour per day, but more questionable claims and articles would have been found if they had spent more time.
JCEJ sent out daily emails containing questionable social media posts gathered by students to 19 collaborating journalists at national and local newspapers as well as TV broadcasts. The journalists received a total of 195 questionable social media content and posts, excluding incorrect reports by major media outlets and remarks by politicians that we did not intend to fact check in the project. JCEJ then posted a total of five items debunked by three or more journalists on its blog.
The following chart shows details of the five pieces of content labeled as “debunked（フェイク判定）”, their original sources, and which platforms they were mainly disseminated.
How false content on opposition and ruling parties spread differently
Four of the five items debunked by three or more journalists were related to opposition parties. However, when observing other questionable items which we did not post on the JCEJ blog, we noticed that there was also a fair amount of dubious claims associated with the ruling Liberal Democratic Party (LDP) led by Prime Minister Abe. In fact, 47 items were related to the LDP, 45 to the Party of Hope, 27 to the main opposition Constitutional Democratic Party of Japan, 19 to the Communist Party, 16 to the now-disbanded Democratic Party, 6 to the Japan Innovation Party, and 3 to the Komeito party, the LDP coalition partner.
By candidate, 33 were questionable posts related to Prime Minister Abe, 27 to Tokyo Gov. Yuriko Koike, the former leader of the Party of Hope, 9 to the Constitutional Democratic Party of Japan leader Yukio Edano, and 6 to Kiyomi Tsujimoto of the main opposition party.
It is worth mentioning that although we had more questionable social media content related to the ruling LDP and Prime Minister Abe, the debunked items that circulated widely online were mostly associated with opposition parties. For example, there were only six questionable posts and articles collected about the main opposition party’s Tsujimoto –- the former senior deputy secretary-general of the now-disbanded Democratic Party of Japan -- but two of them were labeled as fake.
Although we are conducting further research on this point, fake content related to the ruling party may generate less engagement among internet users compared to that of opposition parties. For example, an article we debunked by J-CAST News, falsely claimed both in its title and content that Tsujimoto went insane and lost her mind during the election campaign period, which was carried by Yahoo News Japan, the country’s biggest news portal.
The false article was then further disseminated through other websites and news aggregation websites. Meanwhile, the spread of political rumors about the ruling LDP and its candidates seem to have been limited within Twitter. Japan’s online media tend to be energized by rumors surrounding opposition parties and its politicians.
Challenges in the debunking process
We only published five debunks by the end of the project. This is partly because the collaborating journalists considered the debunking process fairly time-consuming and complex. For example, in attempting to debunk the article which falsely claimed that Tsujimoto had applied for an official recognition by the Party of Hope, it was not enough to verify with Tsujimoto herself. Very often politicians do not provide journalists the truth, while political situations could develop faster without the candidate knowing about it. Debunking rumors requires journalists to contact multiple sources before reaching accurate judgments.
Despite journalists having to engage in vast amount of work during the debunking process, debunked content is often not considered newsworthy in Japan. One of the participating journalists admitted that “(False news that circulates online) are beneath our notice in light of the existing criteria for what makes news. It’s hard for us to generate stories on them.” In fact, although some media outlets were interested in our project, no newspapers or TV broadcasts directly mentioned or reported the actual content of any of the five debunked items.
Meanwhile, online platforms were disseminating mis- and disinformation about certain political candidates during the election campaign period. This is due to irresponsible internet companies who claim being simply platforms, but not news media. Fake news continues to be disseminated by online media, but neglected by traditional media.
A “cluster” of accounts connected through fake content and filter bubbles
“That politician is a satellite of Xi Jinping.” “Half of the politicians of the four opposition parties are from South and North Koreas.” These were a typical of questionable social media claims we saw.
Some journalists who participated in our project were astonished to see such claims being shared widely as these types of social posts would usually not be found on their timelines. Others played it down, saying nobody would believe such rubbish.
The journalists asked if younger generations such as the students who monitored social media throughout the election campaign would often encounter these types of rumors, but none of the students had never seen such information on their timelines either. “This would drive me crazy if I kept watching (the posts),” said one of the students.
We also found that there are clusters of Twitter accounts that actively share and retweet articles on websites that carry fake news. None of us knew about these accounts before the project, because neither journalists nor students had followed these accounts: filter bubbles had most likely prevented us from spotting them.
* * *
Disclaimer: This blog post is based on the article published in the 2018 February issue of GALAC (Housou-Hihyou-Kondankai), “Fake news surrounding the lower house election: measures to combat disinformation and the challenges.” The original article in Japanese was edited and translated by Hiroyuki Fujishiro and Kayo Mimizuka.
Our project was featured in articles by Nieman Lab and The Straits Times (via Kyodo News).