43% of the people in Taiwan use AI! They are unusually confident in media literacy, but only 10% of the public frequently fact-checks.

The 2025 Taiwan Internet Report survey shows that 43.19% of the Taiwanese population has used AI. While they have high confidence in information recognition, only one in ten people regularly fact-check, indicating a risk of false confidence.

Taiwanese people's AI literacy has improved, leading to a shift in news sources towards social media.

Have you ever encountered someone in a community who takes AI responses as correct and just throws them out, even though the content is riddled with errors? There might actually be quite a few people like that!

The Taiwan Network Information Center (TWNIC) recently announced the 2025 Taiwan Internet Report, revealing an increase in generative AI literacy in Taiwan, and that the public's channels for obtaining news have significantly shifted from traditional television to social media platforms. However, there is a common contradiction in information verification behavior, characterized by “high confidence, low action.”

Generative AI becomes a daily aid, but verification habits have not kept up.

The “Taiwan Internet Report” was conducted between the end of July and early September this year, obtaining 2,142 valid samples with a confidence level of 95%. The survey results show that the Taiwanese public has a high acceptance of emerging technologies, with 43.19% of the population having used generative AI, and 8.54% are willing to pay for subscription services.

Most users show a significant improvement in self-efficacy after crossing the initial usage threshold, indicating that AI has transformed from merely a novel tool into a daily aid for problem-solving.

Source: 2025 Taiwan Internet Report shows that the Taiwanese public has a high acceptance of emerging technologies, with 43.19% of the population already using generative AI.

However, the main channel for Taiwanese people to access news is undergoing a structural change. The percentage of television news has dropped from 39.57% in 2024 to 32.02%, losing 7.55 percentage points in a single year, marking the largest decline in recent years. Social media has risen to become the second-largest news source, reaching 21.11%, an increase of 6.28 percentage points from 2024.

In such an information environment, **as many as 57.11% of internet users are confident in their ability to verify the truth of news or information, **but in reality, more than 60% (65.91%) of internet users never or rarely actively verify information, while those who frequently verify information only account for 10% (13.6%).

Source: 2025 Taiwan Internet Report 57.11% of internet users believe they have confidence in their ability to verify the truthfulness of news or information, but in reality, over 60% (65.91%) of internet users rarely or never actively verify information.

Source: 2025 Taiwan Internet Report. 57.11% of internet users consider themselves confident in their ability to verify the truth of news or information, but in reality, over 60% (65.91%) of internet users rarely or never actively fact-check information.

The people of Taiwan have a false sense of confidence in AI, which may become a pain point for fraud prevention.

In response to the public's high confidence in AI, the Director of the Ministry of Education's Digital Promotion Program, Xu Wenwei, pointed out that nearly 80% of the public希望 the government provides AI literacy education, which instead reflects a phenomenon of “false confidence in AI.”

He suggests that future education should distinguish between two levels: first, basic AI literacy, enabling users to understand the biases and limitations of AI; second, AI protection capabilities, covering fraud prevention, misinformation identification, and personal data protection.

When fraud techniques combine with AI technology and reach a high level of professionalism, Taiwan needs greater investment in education.

Professor Zhang Yumin from the College of Communication at National Chengchi University further analyzed that about 70% of the public believes they can distinguish AI-generated content, but international empirical research shows that the accuracy rate for distinguishing real human photos from AI-generated photos is only 61.3%. Even the most sensitive human facial features are difficult to distinguish between genuine and fake.

Zhang Yumin believes that the Taiwanese public may have an overly positive self-perception, and the only effective solution at the moment is to require platforms to proactively disclose whether the content is AI-generated.

Source: Professor Zhang Yumin from the School of Communication at National Chengchi University believes that the Taiwanese public may have an overly positive self-perception. Currently, the only effective solution is to require platforms to proactively disclose whether content is AI-generated.

The era of generative AI has arrived, and media literacy faces new challenges.

In the era of generative AI, media literacy faces new challenges, and even the television stations themselves are not properly overseeing it.

According to a report by the United Daily News, this year, when broadcasting news, the民視新聞台 mistakenly took AI-generated images as real footage, prompting the National Communications Commission (NCC) to launch revised AI guidelines for broadcast media.

NCC代理主委陳崇樹表示，If a news station uses AI-generated images, it must disclose that the images are AI-generated, and the news station must also fact-check.

The Lonely Crowd in the Digital Age: The Hidden Risks of Highly Educated Groups

With the development of social media platforms and digital tools, it seems that there is an endless reach of the internet, yet a group of “lonely crowds” has emerged.

Wu Qi-yin, a researcher at the Institute of Sociology of the Academia Sinica, pointed out that this is a group of people who rely heavily on digital tools but lack the ability to identify risks, and this group is precisely the highly educated young generation at the core of society.

Wu Qiyin emphasized that the group of people who truly need enhanced education are those who, although skilled in using tools, may become overly confident or lack vigilance, making them more likely to overlook potential risks in the digital wave.