AI Misinformation Rampant: Survey Shows Public Confidence in Identification Ability Increases, Not Decreases
A new survey report from National Taiwan University reveals that despite the increasing prevalence of AI-generated misinformation due to AI's rise, the Taiwanese public's confidence in their ability to identify misinformation has surprisingly increased. This trend contrasts with international observations and warrants further research.
📋 Article Processing Timeline
- 📰 Published: April 30, 2026 at 20:35
- 🔍 Collected: April 30, 2026 at 21:01 (26 min after Published)
- 🤖 AI Analyzed: May 1, 2026 at 10:05 (13h 3m after Collected)
Central News Agency
(Central News Agency reporter Chen Chieh-Ling, Taipei, 30th) With the rise of artificial intelligence (AI), distinguishing between true and false information has become difficult. National Taiwan University today released its latest major misinformation survey report, showing that the AI user population has increased to over 70%, and the proportion of people receiving AI-generated misinformation has risen. However, the public's self-perceived ability to identify misinformation has not decreased but increased.
The 2025 annual misinformation survey is the fourth year of a series of surveys, conducted by Professor Hung Chen-Ling and Director Hsieh Chi-Lung of National Taiwan University's Graduate Institute of Journalism, and Researcher Shen Chih-Hsin of Academia Sinica. As in previous years, it commissioned the National Taiwan University Dynamic Public Opinion Research Platform to conduct an online survey. The target audience was residents registered in Taiwan, aged 20 and above, surveyed from December 20 to December 25, 2025, completing 2755 valid online questionnaires.
At a 95% confidence level, the sampling error was approximately plus or minus 1.87 percentage points. Hsieh Chi-Lung explained at today's press conference that this survey shows that misinformation has shifted from an occasional phenomenon to a characteristic of the regular information environment, with over 90% of the public generally exposed to it, and even appearing frequently in daily life. At the same time, 96.81% of respondents believe that misinformation has a serious impact on society, which is higher than the 2024 survey.
The survey found that the most common misinformation content encountered by the public focuses on financial investment fraud and political issues. In terms of source judgment, 94.77% believe that fraud groups frequently create misinformation. Hsieh Chi-Lung said that the most common way for people to confirm misinformation is to use their own knowledge and judgment. It is noteworthy that 33.54% of respondents still judge the authenticity of information based on emotions and preferences.
Notably, Hung Chen-Ling said that given the rapid popularization of AI, the proportion of people receiving misinformation has increased, and data has been collected for two consecutive years. The latest survey found that the AI user population has increased, with 72% sometimes or frequently using it, up from 50% in 2024. The proportion of people who agree that AI provides more objective and accurate information than human-provided information has risen to 40% from 35% in the previous survey.
The survey shows that the proportion of people receiving AI-generated misinformation has increased. Hung Chen-Ling stated that when asked if they could identify AI-generated misinformation, the public believes they have the ability to identify it, with the proportion of those who can frequently identify it increasing from 16% to 22%. Those who can sometimes identify it also increased from 38% to 40%, showing the largest increase.
Hung Chen-Ling believes that the public's rising evaluation of AI information as objective and accurate is a topic that warrants further research. In addition, as the public receives more AI-generated information than in previous years, it used to be said that "seeing is believing" with pictures, but with the advent of the AI era, AI's ability to simulate reality is too strong, not only capable of generating images but also voices. The Taiwanese public's confidence in their ability to identify it also needs to be deeply explored.
Hu Yuan-Hui, Chairman of Public Television Service, pointed out that what surprised him most about this survey was the Taiwanese public's confidence in their ability to identify the authenticity of AI information. Based on overseas experience, facing the increasingly powerful AI generation capabilities, people in other countries have generally become less confident in their ability to distinguish between true and false AI information in recent years. The increasing confidence of the Taiwanese public in identification warrants further research.
In addition, this survey for the first time included the risk of fake news in Line groups. The combined analysis of experimental results showed that groups where "everyone discusses enthusiastically but no one manages" or "most people lurk and don't respond" are most likely to be filled with fake news, with a risk as high as nearly 60%. Its influence even surpasses the source of information or group size. Conversely, groups with regular fact-checking and debunking mechanisms can significantly enhance trust. (Edited by Lung Po-An) 1150430
Choose to stand with facts; every sponsorship you provide is a force to protect press freedom.
Download the Central News Agency's "First-hand News" APP to stay updated with the latest news.
The text, images, and videos on this website may not be reproduced, publicly broadcast, publicly transmitted, or utilized without authorization.
(Central News Agency reporter Chen Chieh-Ling, Taipei, 30th) With the rise of artificial intelligence (AI), distinguishing between true and false information has become difficult. National Taiwan University today released its latest major misinformation survey report, showing that the AI user population has increased to over 70%, and the proportion of people receiving AI-generated misinformation has risen. However, the public's self-perceived ability to identify misinformation has not decreased but increased.
The 2025 annual misinformation survey is the fourth year of a series of surveys, conducted by Professor Hung Chen-Ling and Director Hsieh Chi-Lung of National Taiwan University's Graduate Institute of Journalism, and Researcher Shen Chih-Hsin of Academia Sinica. As in previous years, it commissioned the National Taiwan University Dynamic Public Opinion Research Platform to conduct an online survey. The target audience was residents registered in Taiwan, aged 20 and above, surveyed from December 20 to December 25, 2025, completing 2755 valid online questionnaires.
At a 95% confidence level, the sampling error was approximately plus or minus 1.87 percentage points. Hsieh Chi-Lung explained at today's press conference that this survey shows that misinformation has shifted from an occasional phenomenon to a characteristic of the regular information environment, with over 90% of the public generally exposed to it, and even appearing frequently in daily life. At the same time, 96.81% of respondents believe that misinformation has a serious impact on society, which is higher than the 2024 survey.
The survey found that the most common misinformation content encountered by the public focuses on financial investment fraud and political issues. In terms of source judgment, 94.77% believe that fraud groups frequently create misinformation. Hsieh Chi-Lung said that the most common way for people to confirm misinformation is to use their own knowledge and judgment. It is noteworthy that 33.54% of respondents still judge the authenticity of information based on emotions and preferences.
Notably, Hung Chen-Ling said that given the rapid popularization of AI, the proportion of people receiving misinformation has increased, and data has been collected for two consecutive years. The latest survey found that the AI user population has increased, with 72% sometimes or frequently using it, up from 50% in 2024. The proportion of people who agree that AI provides more objective and accurate information than human-provided information has risen to 40% from 35% in the previous survey.
The survey shows that the proportion of people receiving AI-generated misinformation has increased. Hung Chen-Ling stated that when asked if they could identify AI-generated misinformation, the public believes they have the ability to identify it, with the proportion of those who can frequently identify it increasing from 16% to 22%. Those who can sometimes identify it also increased from 38% to 40%, showing the largest increase.
Hung Chen-Ling believes that the public's rising evaluation of AI information as objective and accurate is a topic that warrants further research. In addition, as the public receives more AI-generated information than in previous years, it used to be said that "seeing is believing" with pictures, but with the advent of the AI era, AI's ability to simulate reality is too strong, not only capable of generating images but also voices. The Taiwanese public's confidence in their ability to identify it also needs to be deeply explored.
Hu Yuan-Hui, Chairman of Public Television Service, pointed out that what surprised him most about this survey was the Taiwanese public's confidence in their ability to identify the authenticity of AI information. Based on overseas experience, facing the increasingly powerful AI generation capabilities, people in other countries have generally become less confident in their ability to distinguish between true and false AI information in recent years. The increasing confidence of the Taiwanese public in identification warrants further research.
In addition, this survey for the first time included the risk of fake news in Line groups. The combined analysis of experimental results showed that groups where "everyone discusses enthusiastically but no one manages" or "most people lurk and don't respond" are most likely to be filled with fake news, with a risk as high as nearly 60%. Its influence even surpasses the source of information or group size. Conversely, groups with regular fact-checking and debunking mechanisms can significantly enhance trust. (Edited by Lung Po-An) 1150430
Choose to stand with facts; every sponsorship you provide is a force to protect press freedom.
Download the Central News Agency's "First-hand News" APP to stay updated with the latest news.
The text, images, and videos on this website may not be reproduced, publicly broadcast, publicly transmitted, or utilized without authorization.