On the 15th, the Korea Internet Self-Governance Organization (KISO), which includes domestic portals and AI chatbot companies such as Naver, Kakao, and SK Communications (hereinafter SK Comms), announced that it has begun discussions to respond to deepfakes.
Korea Internet Self-Governance Organization (KISO) CI [Image=KISO]
The New Technology Committee under KISO recently held a meeting to share the current response status of member companies regarding the misuse of AI technologies such as deepfakes and discussed future countermeasures. The New Technology Committee, composed of academic experts, includes major domestic portals like Naver and Kakao, as well as chatbot service providers such as Simsimi, Scatter Lab, and Tunip.
The New Technology Committee is particularly closely monitoring the potential spread of false and manipulated information using deepfake technology ahead of the April parliamentary elections, deciding to collect and analyze specific cases. After gathering user opinions through surveys, they plan to establish appropriate countermeasures from the perspective of self-regulation.
KISO plans to cooperate with the National Election Commission to ensure election fairness by promptly taking action when false and manipulated information using deepfake videos is clearly confirmed.
Naver has established a system on the main page of its reporting center with new reporting sections for ▲election-related false information and ▲deepfake creations, enabling users to filter and respond as much as possible. When videos are posted on blogs, cafes, etc., or when deepfake-related search terms are entered, cautionary and guidance messages regarding the use of deepfakes are provided to protect users. Additionally, in its news service, AI-generated articles are clearly marked within the article body to make users aware of this fact.
Kakao has developed policies and technical measures to help users easily identify AI-generated content themselves. Kakao displays the phrase "This article was generated by AI" at the top of articles created using AI by media outlets so that users can easily recognize it. The generative AI image creation model "Karlo" has introduced technology to embed invisible watermarks. Furthermore, through campaigns using keyword searches and banners, Kakao plans to raise user awareness and provide detailed guidance on how to report deepfakes.
SK Comms plans to prevent the creation of election-related deepfakes by implementing "search term restrictions" concerning the names of major politicians when generating images using its "AI Chat" service. It also intends to operate a reporting channel for election-related false information and display cautionary messages about deepfakes within the service to alert users.
Lee Jae-shin, chairman of KISO’s New Technology Committee, stated, "If false and manipulated information that can be mistaken for facts spreads in the online environment, it will undermine social trust and make it even more difficult to verify the authenticity of countless contents." He added, "We will promote research to respond to social issues caused by the misuse of AI technology."
© The Asia Business Daily(www.asiae.co.kr). All rights reserved.
