[EAI Fake News Conference] Korean-Style Fake News

YouTube Link: https://www.youtube.com/watch?v=d3fQJHvFdFA

Editor's Note

Professor Han Gyu-seop of Seoul National University analyzes that, unlike in foreign countries, fake news in Korea is mostly disseminated after being picked up by political circles and other media outlets, leading to amplification. Based on an analysis of mediation requests and recommendations received by the Korea Media Arbitration Commission, Professor Han revealed that major media companies such as MBC, KBS, and the Chosun Ilbo group have become subjects and channels for the dissemination of misinformation. He points out that this is part of the mainstream media's self-preservation efforts to survive in a competitive media environment centered on portal sites.

f2e34dd1a320d063

Download: Presentation Transcript | Presentation Slides

Presentation Transcript

Han Gyu-seop (Professor, Seoul National University): While other speakers have focused on the recipient side of fake news, my presentation today will address issues related to the supply side. I have titled my presentation "Korean-Style Fake News." As many of you know, the term "fake news" itself is quite ambiguous and often used with overlapping meanings.

I recall that the term "fake news" originated from the English term "Fake news" and was subsequently adopted into Korean. Originally, "fake news" primarily referred to acts close to fraud, where content was created in the format of news articles by media outlets for financial gain, disseminated through social media, and monetized through advertising based on viewership. As these types of schemes became less effective, terms like "misinformation" and "disinformation," which imply more intent, came into use. Misinformation is generally understood as false or manipulated information disseminated by individuals or groups through social media, even if not in an article format.

In the Korean context, what we commonly refer to as "fake news" seems to manifest differently from these two general definitions used internationally. When you recall common examples of what is called "fake news" in Korea, they often do not fit either of the definitions I just mentioned. Instead, most of what is called "fake news" in Korea appears to be content originally produced as formal articles by officially registered media organizations, which then gets amplified after being picked up by political circles or other media outlets.

Among what we recognize as misinformation, the "Cheongdam-dong late-night drinking party incident" is perhaps the most representative example of what is called fake news in Korea. This incident was first reported by a media outlet called <The Tamsa> and then widely covered by established media companies, leading to its amplification. Similarly, the report on the Kim Man-bae audiotape was initially broadcast by <Newstapa> and subsequently amplified by political circles and other media outlets.

Looking at the personnel composition of media outlets like <The Tamsa> and <Newstapa>, most of their staff come from backgrounds in established media. Therefore, in terms of personnel, these outlets are not significantly different from established, mainstream media. In conclusion, the fact that "Korean-style fake news" mostly refers to media articles reporting unverified facts is a stark contrast to foreign cases. Consequently, it is very difficult to approach the core of the issue by conducting research in Korea based on the definitions and research frameworks of fake news developed overseas.

Therefore, to understand the reality of misinformation dissemination, I believe it is crucial to first examine the spread of misinformation through media outlets. The East Asia Institute (EAI) survey also revealed results from a questionnaire indicating that voters already share the perception that a significant amount of fake news originates from media channels. The survey included a question: "How much responsibility do you believe the following individuals/groups have for the production and dissemination of fake news or false information?" As Professor Yoon Sung-hee presented, while YouTubers (both conservative and progressive) were identified as having the highest responsibility, the next highest were conservative and progressive media outlets, with 56.4% and 55.4% respectively.

This indicates that media outlets are identified as a cause of fake news dissemination at a statistically significant level, even higher than ruling and opposition party politicians. This situation is quite different from cases in other countries.

So, who is responsible for this media-driven misinformation? More specifically, we need to consider whether mainstream media or online/new media outlets bear greater responsibility.

From a conventional perspective, one might assume that newly established online media or non-mainstream outlets are more responsible. This view often holds that mainstream media, with their adherence to journalistic traditions, internal training, and gatekeeping mechanisms, are functioning to some extent.

However, I believe there are several reasons, as mentioned by Professor Kang Won-taek and others, to suggest that mainstream media may have the same or even more severe problems. Given the extreme political polarization among voters, as shown by Professor Kang, they are particularly vulnerable to misinformation. The data collected in this study shows that even for the "Cheongdam-dong late-night drinking party incident," where it was clearly established that the event did not occur through reports from numerous media outlets, 52.5% of respondents still believed that President Yoon and 30 lawyers from Kim & Chang had a late-night drinking session at the Cheongdam-dong establishment.

Another aspect is the significant polarization within the media itself, comparable to that of the voters. Particularly in South Korea, public broadcasting, which has a fragile governance structure and is often dominated by specific factions, experiences intense inter-factional conflict, creating a media environment vastly different from other countries. Furthermore, some conservative media outlets are strongly ideologized, providing a strong incentive for these media organizations to disseminate misinformation.

Finally, the media ecosystem has largely collapsed due to the dominance of portal sites in news distribution. This has led to an exponential increase in the number of media outlets, with over 10,000 registered media companies in South Korea, resulting in fierce, cutthroat competition. Since news inherently deals with daily events, intensifying competition makes it extremely difficult to produce higher-quality content. Consequently, the media industry has increasingly focused on easily implementable strategies like ideological marketing or coercive sponsorship deals with corporations. This economic environment significantly increases the likelihood of media-driven misinformation, and even established mainstream media with high operational costs are not immune to these pressures.

In my analysis, I have compiled and analyzed mediation requests submitted to the Korea Media Arbitration Commission (KMAC) from 2017 to May 2023. A mediation request implies a potential error in the article, and a recommendation for mediation suggests that the applicant's claim has some basis, thus indicating the presence of misinformation in the article. Therefore, I have interpreted these as indicators of misinformation.

In this sense, this data can be considered the most objective measure for examining the dissemination of media-driven misinformation. During this period, mediation requests were filed for 23,367 articles written by 2,164 media outlets. This amounts to approximately 3,600 mediation requests annually, or about 10 per day, submitted to the KMAC. Of these, 35%, or about 8,280 cases, resulted in a recommendation for mediation. This means that approximately three to four articles per day are recommended for mediation, suggesting that three to four articles containing misinformation are disseminated daily. However, it is important to note that many victims of misinformation face significant burdens when considering mediation requests or legal action against media outlets due to the potential for retaliatory reporting.

Therefore, it is reasonable to assume that the actual number of articles containing misinformation, and the number of hidden victims, is likely much higher than these figures suggest.

Looking at the data by media outlet, the list on the left shows the top 30 media outlets that received the most mediation requests during this period. MBC, KBS, imbc, SBS, Chosun.com, and JoongAng Ilbo Online were among those with the highest number of mediation requests. This list reveals several interesting facts: prominent media outlets, including public broadcasters, terrestrial broadcasters, Chosun Ilbo, and JoongAng Ilbo, are ranked at the top, not just online-only media.

Furthermore, MBC ranked first with 650 cases, and its subsidiary imbc ranked third with 512 cases, raising the question of whether these two entities should be considered separately. Similarly, while Chosun Ilbo itself is not listed, its subsidiaries such as Chosun.com and JoongAng Ilbo Online, which produce online articles for prominent media groups, are all ranked at the top. This likely reflects how major media outlets are responding to the market conditions in Korea, as previously discussed.

Established major media outlets appear to be creating subsidiaries that exclusively produce online articles, disseminating content that is more sensational and contains unverified information than their print counterparts. Including articles from these subsidiary-like companies, the table on the right shows that MBC and KBS still ranked first and second, followed by Chosun Ilbo, SBS, YTN, JTBC, JoongAng Ilbo, Hankyoreh, Kyunghyang Shinmun, Yonhap News, and all other well-known media outlets within the top 10.

This phenomenon, where public broadcasters and major newspapers rank highest in the dissemination of misinformation, is likely unique to Korea. Analyzing the cases where mediation recommendations were actually issued yields similar conclusions. MBC, KBS, and Chosun.com are among the top-ranked outlets. When subsidiary companies are included, Chosun Ilbo, KBS, and MBC received significantly more mediation recommendations compared to other media outlets.

For a more rigorous analysis, I employed a Bayesian hierarchical model to analyze annual mediation requests and recommendations for each media outlet. The fixed effects considered in this model include variables that account for all other media outlets, primarily internet-based ones, as well as public broadcasters. The figures on the right indicate how much more likely these media outlets are to receive mediation requests compared to the baseline. The random effects estimates on the right also show, as observed previously, that MBC, KBS, imbc, Internet KBS, and Chosun.com received the highest number of mediation requests.

Public broadcasters and major media outlets ranked highest. Another interesting observation is that among major media outlets, the "닷컴" (dot-com) versions of newspapers like Chosun Ilbo, JoongAng Ilbo, Hankyoreh, and Dong-A Ilbo received more mediation requests than their print counterparts. The only exceptions are KBS and MBC, which are characterized by strong ideological biases, allowing for more assertive claims within their broadcasts. When all affiliated companies are aggregated, KBS, MBC, and Chosun Ilbo emerge as the "big three," with Dong-A Ilbo and Maeil Business Newspaper also showing higher mediation request rates than the baseline internet media.

Analyzing cases that received mediation recommendations also reveals a similar pattern, with KBS and MBC having overwhelmingly more cases, followed by Chosun.com, imbc, and Internet KBS. Among major media outlets, the "닷컴" versions received more mediation recommendations than their print counterparts. When analyzed collectively, MBC, KBS, and Chosun Ilbo are the "big three." Interestingly, while Naver is often criticized as a breeding ground for fake news due to its significant role in news distribution, it received considerably fewer mediation recommendations compared to general internet media or major outlets. Therefore, labeling Naver as the sole source of fake news appears inaccurate.

Subsequently, many media companies may receive mediation requests or recommendations. If they receive many mediation requests, the mediation recommendations are bound to increase. Thus, modeling the probability of a mediation recommendation being issued among the received mediation requests reveals a slightly different picture. In terms of probability, Hankyoreh TV showed the highest likelihood. As Hankyoreh TV did not receive a particularly large number of mediation requests, it may not be possible to view this as a problem in itself. Rather, this might be more meaningful when viewed collectively.

When analyzed collectively, TV Chosun showed the highest probability, followed by Hankyoreh Shinmun, KBS, Chosun Ilbo, and Dong-A Ilbo.

To summarize the results, due to the characteristics of the Korean media environment, it can be concluded that not internet media companies, but rather public broadcasters and mainstream media companies with strong political leanings are generating the most media-driven misinformation. MBC, KBS, and Chosun Ilbo-affiliated media companies showed the highest frequency in terms of mediation requests and recommendations. A significant number of major media companies have established subsidiary .com companies that produce articles distributed online, which are becoming a major channel for misinformation distribution. Thank you. ■

■ Han Gyu-seop_Professor, Department of Media and Information, Seoul National University.

■ Editor: Kim Sun-hee, Senior Researcher, EAI, Park Ji-soo, Researcher, EAI

Contact: 02 2277 1683 (ext. 208) | jspark@eai.or.kr

Video Transcript

Hello, everyone. It's 10:50 AM, so I'm not sure if everyone is still alert. Regardless, I'm not sure if my presentation today will align well with what other speakers have discussed. While others may have focused more on the recipient side or similar aspects of fake news, I will be addressing issues related to the supply side. My presentation title, which I hope is visible, is "Korean-Style Fake News." As most of you are aware, the term "fake news" is quite ambiguous and often used with overlapping meanings. I recall that the term "fake news" originated from the English term "Fake news" and was subsequently adopted into Korean.

Originally, "fake news" primarily referred to acts close to fraud, where content was created in the format of news articles by media outlets for financial gain, disseminated through social media, and monetized through advertising based on viewership. As these types of schemes became less effective, terms like "misinformation" and "disinformation," which imply more intent, came into use. Misinformation is generally understood as false or manipulated information disseminated by individuals or groups through social media, even if not in an article format. However, in the Korean context, what we commonly refer to as "fake news" seems to manifest differently from these two general definitions used internationally. When you recall common examples of what is called "fake news" in Korea, they often do not fit either of the definitions I just mentioned. Instead, most of what is called "fake news" in Korea appears to be content originally produced as formal articles by officially registered media organizations, which then gets amplified after being picked up by political circles or other media outlets.

While many people may consider certain information to be false, among what we recognize as misinformation, the "Cheongdam-dong late-night drinking party incident" is perhaps the most representative example of what is called fake news in Korea. This incident was first reported by a media outlet called <The Tamsa> and then widely covered by established media companies, leading to its amplification. Similarly, the report on the Kim Man-bae and Shin Hak-rim audiotape was initially broadcast by <Newstapa> and subsequently amplified by political circles and other media outlets. Looking at the personnel composition of media outlets like <The Tamsa> and <Newstapa>, most of their staff come from backgrounds in established media. Therefore, in terms of personnel, these outlets are not significantly different from established, mainstream media.

In conclusion, the fact that "Korean-style fake news" mostly refers to media articles reporting unverified facts is a stark contrast to foreign cases. Consequently, it is very difficult to approach the core of the issue by conducting research in Korea based on the definitions and research frameworks of fake news developed overseas. Therefore, to understand the reality of misinformation dissemination, I believe it is crucial to first examine the spread of misinformation through media outlets. The East Asia Institute (EAI) survey also revealed results from a questionnaire indicating that voters already share the perception that a significant amount of fake news originates from media channels. The survey included a question: "How much responsibility do you believe the following individuals/groups have for the production and dissemination of fake news or false information?" As Professor Yoon Sung-hee presented, while YouTubers (both conservative and progressive) were identified as having the highest responsibility, the next highest were conservative and progressive media outlets, with 56.4% and 55.4% respectively.

This indicates that media outlets are identified as a cause of fake news dissemination at a statistically significant level, even higher than ruling and opposition party politicians. This situation is quite different from cases in other countries. So, who is responsible for this media-driven misinformation? More specifically, we need to consider whether mainstream media or online/new media outlets bear greater responsibility. From a conventional perspective, one might assume that newly established online media or non-mainstream outlets are more responsible. This view often holds that mainstream media, with their adherence to journalistic traditions, internal training, and gatekeeping mechanisms, are functioning to some extent. However, I believe there are several reasons, as mentioned by Professor Kang Won-taek and others, to suggest that mainstream media may have the same or even more severe problems. Given the extreme political polarization among voters, as shown by Professor Kang, they are particularly vulnerable to misinformation. The data collected in this study shows that even for the "Cheongdam-dong late-night drinking party incident," where it was clearly established that the event did not occur through reports from numerous media outlets, 52.5% of respondents still believed that President Yoon and 30 lawyers from Kim & Chang had a late-night drinking session at the Cheongdam-dong establishment.

Another aspect is the significant polarization within the media itself, comparable to that of the voters. Particularly in South Korea, public broadcasting, which has a fragile governance structure and is often dominated by specific factions, experiences intense inter-factional conflict, creating a media environment vastly different from other countries. Furthermore, some conservative media outlets are strongly ideologized, providing a strong incentive for these media organizations to disseminate misinformation. Finally, the media ecosystem has largely collapsed due to the dominance of portal sites in news distribution. This has led to an exponential increase in the number of media outlets, with over 10,000 registered media companies in South Korea, resulting in fierce, cutthroat competition. Since news inherently deals with daily events, intensifying competition makes it extremely difficult to produce higher-quality content. Consequently, the media industry has increasingly focused on easily implementable strategies like ideological marketing or coercive sponsorship deals with corporations. This economic environment significantly increases the likelihood of media-driven misinformation, and even established mainstream media with high operational costs are not immune to these pressures.

In my analysis, I have compiled and analyzed mediation requests submitted to the Korea Media Arbitration Commission (KMAC) from 2017 to May 2023. A mediation request implies a potential error in the article, and a recommendation for mediation suggests that the applicant's claim has some basis, thus indicating the presence of misinformation in the article. Therefore, I have interpreted these as indicators of misinformation. In this sense, this data can be considered the most objective measure for examining the dissemination of media-driven misinformation. During this period, mediation requests were filed for 23,367 articles written by 2,164 media outlets. This amounts to approximately 3,600 mediation requests annually, or about 10 per day, submitted to the KMAC. Of these, 35%, or about 8,280 cases, resulted in a recommendation for mediation. This means that approximately three to four articles per day are recommended for mediation, suggesting that three to four articles containing misinformation are disseminated daily. However, it is important to note that many victims of misinformation face significant burdens when considering mediation requests or legal action against media outlets due to the potential for retaliatory reporting.

Therefore, it is reasonable to assume that the actual number of articles containing misinformation, and the number of hidden victims, is likely much higher than these figures suggest. Looking at the data by media outlet, the list on the left shows the top 30 media outlets that received the most mediation requests during this period. MBC, KBS, imbc, SBS, Chosun.com, and JoongAng Ilbo Online were among those with the highest number of mediation requests. This list reveals several interesting facts: prominent media outlets, including public broadcasters, terrestrial broadcasters, Chosun Ilbo, and JoongAng Ilbo, are ranked at the top, not just online-only media. Furthermore, MBC ranked first with 650 cases, and its subsidiary imbc ranked third with 512 cases, raising the question of whether these two entities should be considered separately. Similarly, while Chosun Ilbo itself is not listed, its subsidiaries such as Chosun.com and JoongAng Ilbo Online, which produce online articles for prominent media groups, are all ranked at the top. This likely reflects how major media outlets are responding to the market conditions in Korea, as previously discussed. Established major media outlets appear to be creating subsidiaries that exclusively produce online articles, disseminating content that is more sensational and contains unverified information than their print counterparts. Including articles from these subsidiary-like companies, the table on the right shows that MBC and KBS still ranked first and second, followed by Chosun Ilbo, SBS, YTN, JTBC, JoongAng Ilbo, Hankyoreh, Kyunghyang Shinmun, Yonhap News, and all other well-known media outlets within the top 10.

This phenomenon, where public broadcasters and major newspapers rank highest in the dissemination of misinformation, is likely unique to Korea. Analyzing the cases where mediation recommendations were actually issued yields similar conclusions. MBC, KBS, and Chosun.com are among the top-ranked outlets. When subsidiary companies are included, Chosun Ilbo, KBS, and MBC received significantly more mediation recommendations compared to other media outlets. For a more rigorous analysis, I employed a Bayesian hierarchical model to analyze annual mediation requests and recommendations for each media outlet. The fixed effects considered in this model include variables that account for all other media outlets, primarily internet-based ones, as well as public broadcasters. The figures on the right indicate how much more likely these media outlets are to receive mediation requests compared to the baseline. The random effects estimates on the right also show, as observed previously, that MBC, KBS, imbc, Internet KBS, and Chosun.com received the highest number of mediation requests.

Public broadcasters and major media outlets ranked highest. Another interesting observation is that among major media outlets, the "닷컴" (dot-com) versions of newspapers like Chosun Ilbo, JoongAng Ilbo, Hankyoreh, and Dong-A Ilbo received more mediation requests than their print counterparts. The only exceptions are KBS and MBC, which are characterized by strong ideological biases, allowing for more assertive claims within their broadcasts. When all affiliated companies are aggregated, KBS, MBC, and Chosun Ilbo emerge as the "big three," with Dong-A Ilbo and Maeil Business Newspaper also showing higher mediation request rates than the baseline internet media.

Analyzing cases that received mediation recommendations also reveals a similar pattern, with KBS and MBC having overwhelmingly more cases, followed by Chosun.com, imbc, and Internet KBS. Among major media outlets, the "닷컴" versions received more mediation recommendations than their print counterparts. When analyzed collectively, MBC, KBS, and Chosun Ilbo are the "big three." Interestingly, while Naver is often criticized as a breeding ground for fake news due to its significant role in news distribution, it received considerably fewer mediation recommendations compared to general internet media or major outlets. Therefore, labeling Naver as the sole source of fake news appears inaccurate.

While many media outlets may receive mediation requests, a higher number of requests can naturally lead to more recommendations. Modeling the probability of receiving a mediation recommendation among those that received requests reveals a slightly different picture. In terms of probability, Hankyoreh TV showed the highest rate. Since Hankyoreh TV did not receive an exceptionally high number of mediation requests, this specific finding may not be problematic in itself. However, when analyzed collectively, it becomes more meaningful.

When analyzed collectively, TV Chosun showed the highest probability, followed by Hankyoreh Shinmun, KBS, Chosun Ilbo, and Dong-A Ilbo.

As you can see from the table, MBC and KBS still led, followed by Chosun Ilbo, SBS, YTN, JTBC, JoongAng Ilbo, Hankyoreh, Kyunghyang Shinmun, and Yonhap News, meaning all the media outlets whose names we recognize are concentrated in the top 10. The phenomenon where public broadcasters and the most influential newspapers show high rankings in the distribution of misinformation is likely unique to Korea. Furthermore, an analysis of cases that resulted in recommendations for correction also reaches very similar conclusions.

Here too, MBC, KBS, Chosun Ilbo, and other media outlets are ranked highly. When affiliate companies are combined, Chosun Ilbo, KBS, and MBC received a significantly higher number of correction recommendations compared to other media companies. For a more rigorous analysis, we examined the annual correction recommendations and decisions using a Bayesian hierarchical model. The model on the right, which accounts for fixed effects, includes all other media outlets, mostly domestic ones, and public broadcasters.

The numbers on the right show how many more correction requests these media outlets received compared to the baseline. Looking at the random effect estimates on the right, MBC, KBS, iMBC, and Internet KBS, along with Chosun Ilbo, received the most correction requests, with public broadcasters and influential newspapers ranking highest. Another interesting point is that among influential newspapers, such as Chosun Ilbo, JoongAng Ilbo, Hankyoreh, and Dong-A Ilbo, all received more correction requests than their respective newspaper editions, with the sole exception being KBS and MBC. This might be because these outlets are strongly influenced by partisan logic, allowing them to make stronger claims more freely.

When all affiliate companies are combined, KBS, MBC, and Chosun Ilbo emerge as the top three, followed by Dong-A Ilbo and Maeil Business Newspaper, all of which received more correction requests than the baseline for internet-based media. Analyzing the cases that received correction recommendations, KBS and MBC were overwhelmingly represented, followed by Chosun, iMBC, and Internet KBS. Among influential newspapers, the Hankyoreh received more correction recommendations than its print edition. A comprehensive analysis also shows MBC, KBS, and Chosun Ilbo as the top three. Interestingly, while Naver is often criticized as a hotbed for fake news due to its extensive news distribution, it actually received significantly fewer correction recommendations than general or influential online media outlets. Therefore, labeling Naver as a hotbed for fake news appears to be an inaccurate assertion.

While there may certainly be responsibility for what has been created, if we look at the actual number of mediation cases, Naver has received significantly fewer mediation recommendations than general internet news outlets or prominent media companies. Therefore, the assertion that portals are hotbeds for fake news appears to be an unsubstantiated claim. Furthermore, many media outlets may receive mediation requests or recommendations. If a large number of mediation requests are received, the number of recommendations is bound to increase. Modeling the probability of a mediation recommendation being issued among received mediation requests reveals a somewhat different picture. In terms of probability, Hankyoreh TV showed the highest probability.

In reality, Hankyoreh TV did not receive many mediation requests, so it may not be possible to consider that itself as a problem. It might be more meaningful to view these collectively. When analyzed together, TV Chosun showed the highest probability, followed by Hankyoreh Shinmun, KBS, and Dong-A Ilbo. Summarizing the results, it can be concluded that due to the characteristics of the South Korean media environment, it is not internet news outlets but rather public broadcasters and mainstream media organizations with strong political leanings that are generating the most misinformation. MBC, KBS, and Chosun Ilbo-affiliated media organizations showed the highest frequency in terms of mediation requests and recommendations. Many major media companies have established dot-com subsidiaries that produce articles distributed online, and these are becoming a major channel for the distribution of misinformation. This is likely the extent of it.

Thank you.