In the age of social media, Indigenous peoples in Taiwan continue to grapple with marginalization and a lack of public discourse. Informed by the researcher’s lived experience as a member of an Indigenous community, this article offers an insider perspective on these challenges—and particularly underscores the urgency of accurate representation in the era of artificial intelligence (AI). Without proactive measures, Taiwan faces the risk of Indigenous and national identities being manipulated, leading to harmful distortions of public opinion.
This article explores this issue by examining three examples of disinformation related to the Taiwan Indigenous peoples on the internet and social media from the Taiwan FactCheck Center, established by the Taiwan FactCheck Foundation (TFC, 台灣事實查核中心). The TFC is a non-profit organization established in 2018 by the Association for Quality Journalism and the Taiwan Media Watch Foundation. Their mission is to fact-check claims related to public interest issues, enhance digital literacy, and improve the overall information ecosystem in Taiwan to build a more resilient democratic society. [1]
Example #1: A Supposed Statement of a Member of the Plains Indigenous Peoples about Taiwanese-Japanese 228 Cooperation [2]
The first example is from the fact-check claim: “Who are Japanese Taiwanese? The Chinese Disinformation that Fixated on the Ties between Taiwan and Japan” by Dr. Wei-Ping Li, a research fellow at the Taiwan FactCheck Center. Dr. Li wrote:
Since January 2024, another article of a similar fake statement regarding “Japanese Taiwanese” has also gained traction. The article, titled “A Statement of a Member of the Plains Indigenous Peoples,” told the inaccurate story that the Japanese fought with pro-Japanese Taiwanese against the Chinese Nationalist government during a historical conflict, the “228 incident,” in 1947, hoping to regain control of Taiwan. Although they lost the battle, several Japanese descendants became influential Taiwanese politicians, forming a “New Tide faction” within the Democratic Progressive Party and carrying on the secret mission to “save Japan.”
Dr. Li further explains that the statement “fabricated data and twisted historical events” and “misrepresented recent events to provoke hatred towards Taiwanese and Japanese people.”
What is noteworthy is that the statement was supposedly posted by “a military student” at the Republic of China (ROC) Military Academy in Taiwan. Although the Taiwan Indigenous Peoples currently only make up about 2.5 percent of the overall population in Taiwan, there is a relatively high percentage of Indigenous Peoples joining the military forces in Taiwan. [3] Moreover, the Indigenous peoples are critical in elite units. Therefore, the claim that this was posted by a military student implied that the poster might be Indigenous.
Disinformation such as this that targets concepts such as national identity or ethnic identity is particularly concerning. Attempts to exploit the existence of the Taiwan Indigenous peoples as a manipulable form of identity should be carefully addressed in discussions and policy considerations related to enhancing digital literacy—as well as improving Taiwan’s overall information ecosystem to build a more resilient democratic society, as organizations such as Taiwan FactCheck Center aim to do.
Example #2: A Twitter Account Falsely Claimed that Taiwanese Indigenous People were Killed and Condemned the Government
This second example comes from the TFC post titled “A Twitter Account Spreads Fabricated News about Taiwanese Government’s ‘Wrongdoing’.” This post takes note of “a series of fake news from a certain Twitter account accusing the Taiwanese government of bombing the [I]ndigenous communities, using police force to repress the [I]ndigenous people, and tearing down their houses. The posts used real news as the source but forged the captions.”
The original video featured in the post came from news reports about riots in Honiara, the capital of the Solomon Islands, which occurred in November 2021. The report on FactCheck Center indicated that, in actuality, the videos depicted “violent demonstrations [that] were sparked by the country’s controversial decision to switch diplomatic allegiance from Taiwan to China.” In 2023, the same video circulated on social media platforms and messaging apps, carrying the fake message that an Indigenous person’s building was on fire in Taitung, Taiwan (see image below). The false messaging claimed that this was caused by the government throwing firebombs with drones, with the result that at least 400 people in the Indigenous community were killed. Through this falsified reframing, the disinformation attempted to portray the Taiwanese government as violently attacking its own Indigenous people.
Image: “Taiwan Indigenous Residence Destroyed by the Government.” A social media post that falsely claimed to show the aftermath of the bombing of an Indigenous community by Taiwan’s government; the actual footage was of a 2021 riot in the Solomon Islands. (Image source: Taiwan FactCheck Center)
Example #3: The Claim that a Taiwan Indigenous Tap Dance Troupe Will Embark on a World Tour, Using a Video from Yunnan Province, China
Since May 2024, a four-minute video has been widely circulated on messaging apps and social media platforms. The video features a group of dance performance accompanied by the claim, “Get a Sneak Peek at the Taiwan Indigenous Tap Dance Troupe, Soon to Embark on a World Tour!” However, the video is actually from a Chinese video-sharing platform. It depicts the culture of the Hani people, an ethnic minority in Yunnan, rather than a representation of Taiwan Indigenous peoples or our cultural expressions.
How AI-Generated Content Can Harm Indigenous Cultural Expression
Due to cultural gaps and disruptions caused by successive generations of colonization and government policies, the cultural expressions and/or identities of the Taiwan Indigenous peoples have long been fading away. Disinformation or AI-generated media that features inauthentic depictions of Indigenous identities further contributes to confusion about what cultural signifiers—such as clothes, traditions, language, etc.—relate to each identity, and creates further obstacles for those who are trying to pass on accurate cultural knowledge about those same groups. Even when misinformation about Indigenous groups is not created maliciously, AI-generated images reproduce cultural biases that reinforce misunderstandings about Indigenous people.
One example of this can be seen in a recent AI-generated picture that claimed to depict members of Taiwan’s Kanakanavu people (卡那卡那富族), which was reported in a major Taiwan news outlet (see image below). The accompanying article indicated that the AI-generated image depicted Indigenous peoples of Southeast Asian island nations, rather than authentically reflecting the appearance or cultural expressions of the Taiwan Indigenous peoples. This sparked dissatisfaction within Indigenous communities and the Kanakanavu people, who emphasized that this was a case of misrepresentation. In response, the Bureau of Cultural Affairs explained that the image resulted from biased outputs generated during AI training and was not a case of intentional misuse—thereby pointing out the inherent biases in AI-generated cultural content.
Image: An AI-generated image that inaccurately depicted the Kanakanavu people in Taiwan. The image originated with the Bureau of Cultural Affairs, Kaohsiung City Government, and was reported on by a number of media outlets. (Annotated image source: Pier-2 Art Center, Kaohsiung)
With the growing prevalence of AI-generated content and technologies, it is crucial to approach the future of the Taiwan Indigenous peoples and our representation on social media with caution. This article suggests that proactively addressing these challenges is essential to preventing even more significant difficulties in the future. This effort is not only about ensuring accurate representation for the next generation of Indigenous community members but also about safeguarding Taiwan as a whole from attempts to manipulate national and ethnic identity. Without such measures, the nation could become vulnerable to external forces steering public opinion in harmful directions.
Reimagining the future requires designing mechanisms that allow marginalized communities to use these technologies and actively participate in their development and application. Looking broadly at the use of AI in Taiwan concerning Indigenous peoples, the focus has largely been on its instrumental role in language revitalization and preservation. [4] However, comprehensive policies have yet to systematically review or reflect on how the perspectives of the Taiwanese Indigenous peoples can continue to maintain their visibility and significance in the AI era, both for current and future generations.
Questions and Recommendations Regarding AI Content and Indigenous Communities
At the conclusion of this article, I wish to pose the following questions as key considerations for shaping all relevant policies:
- How might AI and social media influence marginalize Indigenous communities’ cultural preservation and identity construction? And, could these technologies also serve as tools to resist assimilation pressures and preserve cultural diversity?
- How can marginalized communities leverage these technologies to redefine their identities and values, challenging dominant narratives?
- Can social media foster collective action among marginalized communities as a global connecting tool? Could AI help these groups find shared languages and strategies in the context of globalization?
I also urge authorities and relevant policy-making bodies to address the issues that Indigenous peoples face in the social media and AI era, for both the current and future generations. As a potential first step, policymakers could look to international practices for inspiration. For instance, the “Digital Strength Initiative” by Stanford University’s Social Media Lab in the United States, or Canada’s national-level AI policy, which includes courses on “Indigenous Perspectives in AI,” serve as valuable examples. Such initiatives could offer meaningful references for Taiwan’s Indigenous peoples or other marginalized groups who are navigating the challenges of social media and the age of AI.
Envisioning the future should be about adapting to technological advancements and using these technologies to create futures belonging to marginalized communities. This includes transformative changes in economic development, political participation, and educational opportunities. This open-ended and challenging topic offers a valuable opportunity to explore how technology can create a more inclusive and equitable future for marginalized communities. At the same time, it serves as a reminder to maintain critical thinking in the face of technological advancements, in order to prevent further exacerbating inequalities.
The main point: Without proactive measures taken to protect Indigenous communities—and Taiwan as a whole—AI-generated images and disinformation have the potential to manipulate national and ethnic identities through distorted history, falsified current events, and by replicating cultural misunderstandings and biases. Policymakers should look to international AI practices as inspiration, and include Indigenous communities and marginalized groups in their solutions, in order to ensure they have agency over their own representation.
[1] See also Sheera Frenkel et. al., “Surge of Virus Misinformation Stumps Facebook and Twitter,” The New York Times (March 8, 2020; Updated June 1, 2020), https://www.nytimes.com/2020/03/08/technology/coronavirus-misinformation-social-media.html.
[2] This whole statement can be found on “Cofacts,” an online fact-checking bot in Taiwan: https://cofacts.tw/article/3j3lxxl44drog. More about “Cofacts” can be found on: https://en.cofacts.tw/. See also Billion Lee, The Bot Fighting Disinformation: The Story of Cofacts, Taiwan Insight (Oct. 12, 2022), https://taiwaninsight.org/2022/10/12/the-bot-fighting-disinformation-the-story-of-cofacts/.
[3] The Council of Indigenous Peoples, YuanMinHu Icyang ZhuWei Huijian GuoFangBu ShouWei Yuanzhumin FuBuZhang Hsu Yen-Pu ShangJiang (原民會夷將Icyang主委會見國防部首位原住民副部長徐衍璞上將) [The Chairperson of the Council of Indigenous Peoples, Icyang, met with General Hsu Yen-Pu, the first Indigenous Deputy Minister of the Ministry of National Defense] (March 20, 2024), https://www.cip.gov.tw/zh-tw/news/data-list/E93DA6B5E2130657/CAA62483EAB5AEA5877BAABB2CF4DB88-info.html.
The original text is: “原住民志願役軍人共1萬3千多人,占國軍8.7%。原住民族只占臺灣人口數的2.5%,尤其在特勤部隊中更扮演關鍵力量,如「憲兵特勤隊」、陸軍特指部「高空特種勤務中隊」及海軍陸戰隊「兩棲偵搜大隊」等,原住民軍人比率均過半;原住民從軍人數眾多,肩負保國衛民之重任。”
The translation is: “There are over 13,000 Indigenous volunteer soldiers, accounting for 8.7% of the national armed forces, despite Indigenous peoples comprising only 2.5% of Taiwan’s total population. They play a particularly critical role in elite units [emphasis added] such as the ‘Military Police Special Services,’ the Army’s ‘Airborne Special Service,’ and the Marine Corps ‘Amphibious Reconnaissance Battalion,’ where Indigenous soldiers make up more than half of the personnel [emphasis added]. Indigenous soldiers are significant in number and shoulder the important responsibility of defending the nation and protecting its people.”
[4] For example, Hsiao Yi-Hao (蕭一豪), Yuanzhuminzu YuYan RenGongZhiHui KeJi PingTai YanFa Yu JiaoYu YingYong (原住民族語言人工智慧科技平台研發與教育應用) [The Development and Educational Applications of an AI Technology Platform for Indigenous Languages], Taiwan Technology for Inclusion, https://www.ttfi.com.tw/project02-item/aboriginal-language; see also Cemedas Dumalalrath (江子芊) and Kawa (施俊銘), AI KeJi PingTai YanFa YingYong ZuYu XueXi XinJinZhan (AI科技平台研發應用 族語學習新進展) [AI Technology Platform Development and Applications: New Advances in Indigenous Language Learning], Indigenous People Cultural Foundation (April 18, 2024), https://www.ipcf.org.tw/-/News/Detail?newsId=24041814491735522.