[Initial Insight] Measures Needed to Counter 'Deepfake' Shaking Up the Election Scene

Organized Abuse in Taiwan Presidential Election
Awareness Needed to Minimize Side Effects

"Isn't the issue resolved since the amendment to the 'Public Official Election Act,' which bans election campaigning using deepfake videos from the 29th until the general election, passed the National Assembly plenary session?"

As the world faces the 'super election year,' the manipulated technology deepfake, which is fake but looks real, is shaking up the election scene, raising urgent calls for countermeasures in South Korea as well. However, as seen in remarks from a political insider, the political sphere's sense of urgency is not significant. The prevailing complacent perception is that legislative responses have begun with the passage of the Public Official Election Act amendment, so there is no problem. Deepfake technology, which synthesizes images and videos using ever-evolving artificial intelligence (AI) technology to make them appear real, is rampant and even misused in public elections, but awareness is not keeping pace with technological advancements.

On the 13th, Taiwan, which held the world's first presidential election, announced that the 'AI election' had officially begun. Until just before the presidential election vote, Taiwanese voters had to continue fighting against deepfakes. Pro-China misinformation flooded platforms like TikTok, and deepfakes were systematically exploited in election activities, clouding voters' judgment. Bill Gates, founder of Microsoft (MS), expressed concern on his blog last year that "AI-generated deepfakes and misinformation could undermine elections and democracy." It is not just elections. AI technologies, including deepfakes, have been used as sophisticated psychological warfare weapons in the nearly two-year-long Russia-Ukraine war and the Israel-Hamas conflict ongoing since October last year.

Academia warns of the possibility of a deepfake nuclear storm in South Korea as well. It is suggested that deepfakes could emerge as the biggest variable overturning the outcome in the presidential election three years from now, rather than the upcoming general election with over 1,000 candidates. If deepfakes cloud voters' judgment in a presidential election where attention is focused on a few candidates, they could become a decisive factor determining the election result. Considering that in the last 20th presidential election, the vote difference between People Power Party candidate Yoon Seok-yeol and Democratic Party candidate Lee Jae-myung was only 0.73 percentage points?the smallest margin in the history of presidential elections?there can be no complacency regarding the negative influence of deepfakes. Especially since deepfakes, unlike fake news, are difficult even for experts to easily determine as manipulated and require a long time for verification, greater caution is necessary. This is also why U.S. President Biden, ahead of the November presidential election this year, hired hundreds of staff and volunteers to respond to misinformation on social media.

The World Economic Forum (WEF, Davos Forum), where famous political and business figures from around the world gather to discuss common human issues, opened on the 15th (local time) in Switzerland, with AI as the central topic this year. Leaders from various countries are considering ways to utilize AI while reducing side effects such as misinformation production and digital divides. In South Korea, it is necessary to raise awareness of deepfake technology and actively prepare countermeasures. The concept of deepfake election campaigning should be clarified more precisely, and regulatory measures for online platforms should be carefully reviewed and established to minimize side effects.

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