[The Editors' Verdict] Article Arrangement Must Be Made Transparent with 'Filter Bubble' Prevention Law

Bill to Reform Article Arrangement Submitted
"Algorithm Secrecy Encourages External Verification"

At the National Assembly level, there is an expectation that the issue of disclosing the article arrangement algorithms of portal news sites such as Naver and Kakao will be brought into public discussion. This opportunity arose as the 'Partial Amendment to the Act on the Promotion of Newspapers, etc. (Newspaper Act)' proposed by Kim Seung-su, a member of the People Power Party, in April was recently submitted to the relevant standing committee of the National Assembly for formal discussion. Although lawmakers from both ruling and opposition parties, including Jeong Hee-yong and Kim Nam-guk, have previously proposed similar amendments to the Newspaper Act, all were made before the current government took office. Given that the Yoon Seok-yeol administration has pledged to address the issue of confirmation bias caused by recommendation algorithms, Kim’s bill can be seen as reflecting the current administration’s intent.

Kim’s amendment to the Newspaper Act centers on revising the article arrangement by internet news service providers and establishing a separate organization called the Internet News Promotion Committee. This committee would be tasked with providing opinions or recommending corrections regarding arrangement policies and the article supply process.

The demand to disclose article arrangement policies is widely regarded by experts as essential, considering the significance of these platforms. According to the Korea Press Foundation’s '2022 Media Audience Survey,' 75.1% of respondents consume news through internet portals, with portal usage rates among people in their 20s to 40s exceeding 90%, surpassing television usage rates.

However, despite their influence, users know very little about how news exposure is determined. For example, Naver explains its news arrangement principles only in vague terms such as "automatically recommended based on algorithms formed on recency, completeness, and uniqueness." Considering that it is common for articles written exclusively by one media outlet to be overshadowed by similar articles from other outlets, it is difficult to believe that these principles are strictly followed without knowing the specific logic. In short, even if media outlets write and provide articles to portals, they have no way of estimating how these articles are exposed to readers.

The problem with reliance on algorithms stems from the so-called 'filter bubble' phenomenon. A filter bubble refers to the issue arising when internet information providers deliver customized information to users. Despite the political intentions behind news, users may unconsciously be exposed to and influenced by it without awareness. Jin-wook Kim, a lawyer at Juwon Law Firm, pointed out at a recent discussion hosted by the People Power Party’s Portal Committee, "The detailed logic and weighting factors of portal news editing and comment exposure algorithms are not disclosed," adding that "the political inclinations of operators and the temptation for companies to generate unfair profits are structural problems that inevitably get reflected."

There is a prevailing view that algorithm disclosure will ultimately be an unavoidable step for portals. This year, page views on both major portals have sharply declined, leading to assessments that their influence is not what it used to be. Moreover, with Google and Microsoft Bing introducing generative AI as a new weapon, there is no clear countermeasure. For this reason, experts generally agree, with varying degrees, that news arrangement principles should be disclosed. The experts’ evaluations that "portals have brought this on themselves by not disclosing trade secrets" and that "external verification is necessary" carry persuasive weight.

Concerns have been raised that the recent temporary suspension of the News Alliance Evaluation Committee by portals will weaken the self-regulatory function of internet news. However, even during the committee’s activities, quality issues in news persisted. Distrust in algorithms has prevented breaking away from the practice of copying and following others. Transparent disclosure could instead encourage the production of higher-quality content.