As technology advances, convenient generative artificial intelligence (AI) is being developed to produce results based on user needs. Generative AI is gradually advancing in various fields, such as synthesizing voices, creating pictures, and writing novels. However, there are cases where Generative AI is
used unethically, and controversy continues.
For example, techniques for synthesizing and generating images have been used in phishing and sex crimes. People often get caught up in crime because it is difficult to distinguish between real and AIgenerated photos.
These image synthesis techniques also generate fake news and false information. In May 2023, just before voting in Türkiye’s presidential election, a deepfake video spread that “terrorist groups support opposition candidates.” A similar thing happened in Slovakia’s general election in September 2023. Two days before the election, a deepfake voice file was released in which the pro-American opposition leader said, “If our party is to win the election, we need to pay the underprivileged to buy votes.” It was all fabricated, but the files affected voters’ choices.
Due to the dangers of such AI technology, many companies worldwide have decided to limit AIgenerated image creation during elections. The major players in generative AI, Meta, OpenAI, Microsoft, and 17 others, agreed to work collaboratively to prevent deceptive AI content. Midjourney will prevent the creation of political images that include images of politicians such as Donald Trump and Joe Biden by next year. Meta also announced that it will introduce a watermark indicating that it is an AIgenerated image within the next few months.
South Korean platform companies have also developed countermeasures ahead of the 22nd general election scheduled for April this year. Naver displayed a warning message when a user entered a search word related to deepfake production. Kakao Brain is preparing a plan to embed watermarks in AI content.
Elections will be held in about 76 countries around the world this year. The estimated number of
voters is approximately 4.2 billion. When voters check election-related information, they should make sure that the source of the information is reliable. In addition, voters should try not to be deceived by fake AI images. Companies and governments should also do their best to prevent the spread of false information.
