Starting from July 28, the Cultural Heritage Administration will conduct the ‘National Heritage AI Learning Challenge Campaign’ for three weeks, targeting middle school, high school, and university students to correct errors in national heritage information provided by generative AI.
Participants will identify erroneous responses from generative AIs like ChatGPT, learn the correct information, and rectify the results accordingly.
This campaign, introduced for the first time this year, is designed to address repeated issues such as historical inaccuracies, factual distortions, biased descriptions, and inconsistent images in national heritage information provided by generative AIs including ChatGPT, amidst rapidly spreading AI technologies.
For example, to the question, “Who created Hangeul?”, there was an erroneous response generated stating, “King Sejo created the script for the people, based on the Chinese writing system.” Participants will capture such incorrect AI responses, share the corrected information they have instilled in the AI along with their corrected responses on social media, and submit them according to the specified format.
Participation is possible through a campaign-specific QR code, and prizes will be given to outstanding participants.
The Cultural Heritage Administration explained, “Voluntary participation from the public, especially from youth and university students, is crucial in rectifying the distorted presentation of national heritage information by generative AI.”
This activity focuses on highlighting the risks of accepting information provided by AI without scrutiny and creating an opportunity for users to verify the accuracy of the information themselves. The Cultural Heritage Administration plans to continue making policy efforts to ensure that accurate national heritage information is reflected in AI in the future.
