Written by Tommy Hall and Margaret Siu.
Image credit: Image provided by the authors.
When generative, artificial intelligence (AI) programmes are used with malicious intent in the realm of politics, the resulting “deepfakes” can be both convincing (at a glance) and extremely damaging to the target’s image. This article will argue that a unified and proactive response is necessary for Taiwan to successfully navigate the gauntlet of challenges created by AI in a democratic political environment.
The upcoming Taiwanese presidential election already holds examples where false political content created by such technology is released with the intent to manipulate voters. Taiwan FactCheck Center (臺灣事實查核中心), a non-profit and non-governmental organisation dedicated to “making Taiwan a better and stronger democracy” through combatting disinformation, released a public fact-checking report on October 17 that discusses Taiwan’s “first” pre-2024 election, AI-generated fake audio file. In it, AI is used to impersonate the Taiwan People’s Party (TPP) founder and chairman Ko Wen-je (柯文哲) in criticising Vice President Lai Ching-te’s stopover visit to the U.S. and accusing Lai of paying people to attend his events. The controversy emerged in the public spotlight in mid-August when TPP officials sent a report to Taiwan’s Investigation Bureau, vowing to pursue legal action. Ko openly denounced the fake recording’s content, calling the attempt “crude and egregious.” This controversy, one occurrence in a steady stream of computational propaganda targeting Taiwan’s democratic processes, can be seen as an opening salvo for increasingly sophisticated attempts to manufacture disinformation and sway public opinion in a way that benefits the political agendas of those who weaponise emergent AI technology. It warrants a pause for greater concern, as well as a discussion about how computational propaganda is manipulated to influence public opinion in Taiwan and the common threats and challenges it poses to democratic processes across the globe. Ultimately, Taiwan’s democratic resilience will be shaped by its ability to incorporate a constantly evolving, whole-of-society approach to countering all forms of computational propaganda.
Computational Propaganda in a Taiwanese Context
“Computational propaganda” is an umbrella term used by some scholars to describe how social media, artificial intelligence, and big data can work in concert to manipulate public opinion. It is no secret that Taiwan is on the front lines in resisting global attempts to weaponise new technologies against the mechanisms of democracy. China has allegedly allocated 10 billion dollars for its propaganda efforts, and even though the overwhelming majority of this is likely spent to sway domestic audiences, just a tiny drop in this vast propaganda pool could effectively flood the online information channels of its targets abroad—especially Taiwan.
Attempts by China to manipulate Taiwanese public opinion through the island’s domestic media channels did not begin with the advent of social media. China has funded traditional media outlets, such as the WantWant Media Group’s China Times, through steady streams of ad revenue—resulting in coverage shifting to become more China-friendly and leading critics to label the outlet and its leader as being responsible for an infiltration of CCP-backed narratives in the public zeitgeist.
The misinformation landscape in Taiwan’s recent elections shows a complex and constantly evolving challenge. During the 2020 presidential election and the 2021 referendum, disinformation primarily centred on vote rigging, government policies, and cross-strait relations, with unfounded narratives suggesting collusion between the Chinese government and the KMT party. In contrast, the 2022 local elections witnessed a diversification of disinformation topics, extending to national defence and scepticism toward the U.S., influenced by geopolitical events such as Nancy Pelosi’s visit to Taiwan. Additionally, the COVID-19 pandemic became a target for misinformation campaigns aimed at discrediting political figures, exemplified by attacks against Taipei mayor candidate Chen Shih-chung. This adaptive nature of misinformation strategies, which also saw a decrease in direct misinformation on sensitive topics like cross-strait relations, indicates the complexity of the sources and motives behind these campaigns. It is essential to recognise that both foreign, state-affiliated actors and domestic sources contribute to the spread of online misinformation in Taiwan, as evidenced by a misleading story on Taiwan’s Professional Technology Temple (PTT) forum in 2018 about the Chinese Embassy’s supposed rescue efforts in Japan, later attributed to a Taiwanese university student. This multifaceted situation underscores the need for continual vigilance and strategic countermeasures to safeguard the integrity of democratic processes in Taiwan.
Building Taiwan’s Future Resilience: Evolving Solutions to Evolving Technology
Nevertheless, the Taiwanese public has shown increasing awareness and response to misinformation. The 2023 Annual Misinformation Survey by the Taiwan FactCheck Center indicates that a significant majority of Taiwanese citizens are not only encountering misinformation but actively using fact-checking tools for verification. This proactive approach is crucial in the digital era and offers a contrast to misinformation-handling trends in other countries, where direct investigation of information is more common. Nevertheless, Taiwan’s recent elections display continuous shifts in misinformation tactics. Currently, misinformation is used to amplify old political controversies related to food imports, such as pork and eggs. The broadening range of topics, re-circulation of old rumours, and increased use of text-based, fabricated content indicate a sophisticated, evolving approach that threatens the stability of Taiwan’s democracy.
The forthcoming 2024 presidential election in Taiwan stands as a pivotal juncture, with significant implications for the island’s geopolitical stance, especially in its relations with Beijing. This election, marking the end of President Tsai Ing-wen’s term, is expected to be a critical contest between the Democratic Progressive Party, advocating resistance against authoritarian influences, and the Kuomintang, known for its pro-Beijing alignment. The emergence of other strong opposition parties, independent candidates, and the talks of merging opposition tickets make it an increasingly complex political battleground. In this context, the electorate faces an unprecedented onslaught of digital interference. Notably, the spectre of an estimated 100,000 hackers, presumably aligned with Chinese interests, looms large, expecting to zero in on voters in Taiwan’s election. With the pervasive role of social media in Taiwan’s information ecosystem—where approximately three-quarters of the populace engage for news—the digital domain emerges as a highly contested electoral battlefield. The utilitarian application of AI technologies in these campaigns represents a formidable force multiplier, potentially enhancing the efficacy and reach of disinformation efforts significantly.
The role of AI in shaping political narratives is witnessing rapid and expansive growth. This trend, as exemplified by the deployment of AI-generated political advertisements by entities like the Republican National Committee in the United States, foreshadows a similar trajectory in the 2024 Taiwanese presidential election. The implications of this development are profound, potentially altering the landscape of political discourse in ways hitherto unexperienced. In response, entities such as Doublethink Labs and the Taiwan Information Environment Research Centre are spearheading efforts to counteract AI-driven misinformation campaigns. Despite these proactive measures, the challenge posed by AI in the realm of political misinformation remains formidable, necessitating ongoing vigilance and the development of innovative strategies. This battle against digital manipulation in political discourse underscores a broader imperative: the need for robust, adaptive frameworks to safeguard the integrity of democratic processes in an increasingly digitised global political landscape.
In conclusion, the global challenge of AI-driven misinformation, prominently illustrated in the Taiwanese context, necessitates a unified and proactive response. Taiwan’s growing reliance on fact-checking and heightened public awareness exemplify a society adapting to the complexities of the digital era. This situation underscores the imperative for international cooperation in developing robust strategies to combat misinformation. Enhanced digital literacy, transparent AI practices, and adaptable regulatory frameworks are critical in preserving the integrity of democratic processes worldwide. As we navigate this evolving landscape, a collective and vigilant approach is essential for safeguarding democracy in the age of AI.
Tommy Hall graduated from West Point with a B.Sc. in Foreign Languages: Chinese. He is a Marshall Scholar, currently reading for an M.St. in Diplomatic Studies at the University of Oxford. At Oxford, Tommy serves as the Book Review Editor and Opinion Editor for STAIR (St. Antony’s International Review). In 2023, he completed an M.A. in Chinese Studies at SOAS, University of London. Additionally, Tommy is the founder of the Center for Holocaust and Genocide Studies’ Human Rights Initiative (HRI). His peer-reviewed work has previously been published in the Army Cyber Institute’s Cyber Defense Review.
Margaret Siu is a JD candidate at Harvard Law School with a focus on U.S.-East Asian economic statecraft and critical supply chains. She holds a Marshall Scholarship, having earned her M.Sc. at the University of Oxford and the London School of Economics and Political Science. Siu is actively involved in leadership positions with the Harvard International Law Journal and the advisory board of the Oxford Silk Road Society. She is also the founder of Apricity Magazine, an international arts journal. Her notable contributions include work for the UK National Committee on China, Oxford Political Review, and the China Project (SupChina). Siu has prior experience with the Human Rights Initiative at the Center for Strategic and International Studies (CSIS) and in the Trade and Investment team at Taiwan Semiconductor Manufacturing Company’s Corporate and Compliance Legal Division.
This article was published as part of a special issue on Digital Governance in Taiwan and Japan.
