Naver and Kakao Respond to News Bias Criticism with "Algorithm Disclosure Under Review"

[Image source=Yonhap News]

[Asia Economy Reporter Kang Nahum] Naver and Kakao indicated their willingness to disclose their news recommendation algorithms if a social consensus is reached during the 'Portal Algorithm Public Hearing.'

Kim Hee-jung, Head of Kakao's Platform Business Division, stated at the public hearing held by the National Assembly's Science, Technology, Information and Broadcasting Communications Committee, "If discussions regarding algorithm disclosure reach a social consensus, we will actively consider it," adding, "We are internally contemplating this with an open mind to acceptance."

Choi Jae-ho, Director in charge of Naver AiRS, also said, "In 2018, Naver invited professors from computer science and journalism fields to conduct a verification. The current method we use includes the approach suggested by the review committee and is the most reasonable algorithm method in use," adding, "At that time, we explained the source code related to the algorithm implementation stage and method and completed the review."

Before these remarks by the two executives, experts had repeatedly pointed out the need to increase transparency of portal algorithms. Participants in the hearing as witnesses included Lawyer Kim Jin-wook from the law firm Juwon, Kim Dong-won, Policy Director of the Journalists' Union, Professor Choi Kyung-jin of Gachon University, and Professor Lee Soo-young, Emeritus Professor at KAIST.

Lawyer Kim Jin-wook pointed out, "Only the technologies applied to portal news editing and comment exposure algorithms have been disclosed, but detailed logic, weighting factors, or the types of information collected and used have not been revealed, leaving the design and initial value settings of portal AI algorithms still opaque."

Professor Choi Kyung-jin said, "Except for socially clearly agreed lawful standards, it is preferable to avoid mandating the presentation of substantive prior criteria to be reflected in algorithms. Instead, it is necessary to encourage resolution through self-regulation or ethical codes, or prioritize resolution via ex-post regulatory methods or procedural approaches."

However, regarding the level of algorithm disclosure, there were opinions that corporate secrets such as source code should not be disclosed. Professor Lee Soo-young said, "From a developer's perspective, it would be preferable not to disclose source code except in cases of very serious risks."

Policy Director Kim Dong-won also said, "Disclosing source code is difficult, and it would be better to submit transparency reports that journalists or general scholars can use."

This content was produced with the assistance of AI translation services.