Taiwan, widely considered the world’s most-targeted democracy for foreign disinformation, made headlines earlier this month as it became a founding member in a new coalition formed by the International Foundation for Electoral Systems (IFES). Convening policymakers and industry experts, the Artificial Intelligence Advisory Group on Elections (AI AGE) will build bridges between democratic governance and AI expertise to address the challenges and opportunities that AI presents in the conduct of free and fair elections.
Taiwan’s invitation was no courtesy. On the contrary, it has become a laboratory for democratic resilience in the AI age. At a time when those in the pay of authoritarian regimes look to undermine our elections through increasingly sophisticated disinformation campaigns, Taiwan’s playbook is seen as not just effective but essential.
As its peers scramble to address the risks posed by AI, Taiwan has become a sought-after voice in global AI governance. In effect, it has already lived the future others fear, emerging stronger on the other side of AI-fueled disinformation by harnessing the technology’s potential.
So says Taiwan’s Cyber Ambassador Audrey Tang. A headliner at Munich’s Cyber Security Conference and last month’s AI Expo for National Competitiveness in Washington, Tang’s calendar is but one indicator of the importance of Taiwan’s latest key export: its model for digital democracy.
Now the AI AGE offers Taiwan a chance to impart its hard-won lessons, presenting more than 145 countries advised by IFES (including the United States) with opportunities to adopt its best practices. In her remarks at the AI AGE launch, Tang emphasized that Taiwan’s inclusion – as the group’s only member from East Asia – was a reflection of the world’s growing appreciation for its government’s approach to AI. It is “crucial that AI is trained to serve the greater good, while unlocking our collective wisdom and building a more accountable and participatory digital society,” Tang said.
Out of necessity, Taiwan has devised tools that other democracies have not. Other democracies would be wise to adapt Taiwan’s blueprint before it is too late.
Almost 30 years after Taiwan’s first direct presidential election in 1996, new technologies have only emboldened Beijing’s attempts to dismantle the island’s democracy. Its “three warfares” strategy has been adapted to prioritize the use of generative AI (genAI), introducing a more disruptive threat than any of its predecessors. Where traditional models only analyze basic data, genAI is capable of creating original, persuasive content on a massive scale. This makes China’s latest efforts harder to detect and its propaganda far easier to weaponize.
Cyberattacks against Taiwan already posed a significant problem, soaring 30-fold ahead of the 2024 election according to one estimate from Cloudflare. Add genAI to the mix, and the threat level skyrockets. A report from Taiwan’s National Security Bureau to parliament in April revealed it had detected a 60 percent spike in falsified content distributed by China using genAI’s capabilities, from 1.33 million items in 2023 to 2.16 million in 2024.
Countless fake social media posts, articles, newscasts, and even entire books flooded Taiwan’s information ecosystem, with dummy profiles distributing baseless accusations questioning the election’s integrity alongside pro-China and anti-democratic sentiments. The Taiwan FactCheck Center has cataloged a whole host of AI-generated videos intended to cast doubt on the process and its outcomes, including fabricated footage of local politicians and fraudulent images of Xi Jinping commenting on Taiwan’s elections.
The attention paid to Taiwan’s model not only affirms its innovation, but strengthens its international standing despite Beijing’s attempts to isolate it. Its leadership has taken its self-defense seriously, acting quickly but consciously to meet the AI moment.
Taiwan’s tactics are anchored in its core mission, per Tang, to co-govern AI with the people. This ethos has been the guidepost for Taiwan’s response and inspired its blueprint for civic fact-checking, all without compromising its citizens ability to exercise their civil rights. This is significant considering the Taiwanese model employs much of the same tech being leveraged against it by autocrats.
Take Taiwan’s rapid response practices. The Ministry of Education’s False Information Prevention Project implemented curriculum guidelines that equip students with the ability to “pre-bunk” online hoaxes, teaching them to spot and flag fake news. Meanwhile, each of the government’s ministries target attacks en-masse, with engineering teams responding to falsehoods in as quickly as 60 minutes with countervailing narratives. Alongside individual pre-bunking, these officials work to meet their fellow citizens where they are: using AI to their advantage as they share up-to-the-minute graphics, post short videos, and host livestreams to inoculate the public against disinformation.
Then there is Taiwan’s civic-fact checking infrastructure. As journalist Elaine Chan detailed in The Guardian last year, a community of external validators have stepped up to assist the government in its fight for the truth. Groups like the Taiwan FactCheck Center and Doublethink Lab provide citizens with access to needed tools in the pre-bunking phase. China’s operations are formidable, requiring a sizable amount of computing power to counteract. These organizations help meet that need, their presence in turn easing fears of government overreach.
One tool, MyGoPen, shows us how AI can be used in election protection, offering access to a Line account with one-on-one live fact-checking services. Its founder, Charles Yeh, notes that AI “speeds up the checking process – [it] helps with comparison, identification and translation,” drawing on an extensive disinformation database collected by the group. But verification of claims still happens manually; a feat considering its agents were handling up to 3,000 claims a day in the leadup to election day.
Due to these efforts, Taiwan’s elections passed without any major incident. But AI’s faculties will continue to evolve, and democracies must be prepared going forward.
While the kind of sabotage Taiwan faces is distinct, its inclusion in AI AGE and in countless future forums show us that the world’s democrats cannot ignore what Vice President Hsiao Bi-khim has come to call “the AI Island.”
The tech’s full potential has yet to be fully realized; it will continue to advance. Innovators and benefactors, famous and infamous, will continue to fuel the rise of the AI industrial complex. But if the world is serious about securing the future of democracy in the internet age, Taiwan’s resilience must become the norm, not the exception.
