TOKYO - Audrey Tang, Taiwan’s first Digital Minister known for advancing civic participation through technology, visited Japan to share insights on digital democracy. At a Tokyo event, Tang joined University of Tokyo professor Yutaka Matsuo, a leading expert in artificial intelligence, and Katsuya Uenoyama, CEO of AI developer PKSHA Technology, for a wide-ranging discussion on how AI can reshape democratic systems.
The speakers explored a future in which technology empowers inclusive political engagement, enhances decision-making processes, and helps build healthier public discourse in an era dominated by misinformation.
Tang began by explaining how in Taiwan, the Mandarin word for “digital” also implies “plurality,” and that this dual meaning informed the very foundation of her role. Appointed as Digital Minister in 2016, Tang wrote her own job description in poetic form, framing digital technology as a means to enhance shared human experience rather than as a tool of centralized power. “Whenever we hear that a singularity is near,” she read aloud, “let us always remember that plurality is here.”
One of the key examples Tang shared was Taiwan’s response to the spread of deepfake scams. In March of last year, social media platforms such as Facebook and YouTube were flooded with fraudulent advertisements featuring AI-generated likenesses of famous individuals like Nvidia CEO Jensen Huang. These videos promised cryptocurrency or investment opportunities but were entirely fake—yet tech platforms profited from the ads while only removing them after user complaints.
In response, Taiwan’s government sent 200,000 text messages from a trusted public number, inviting citizens to propose ideas on how to deal with the issue. The initiative used language models to summarize and process submissions without distortion, then randomly selected 450 representative participants for a citizens’ assembly held entirely online.
Participants were divided into 45 virtual rooms of 10 people, each moderated by an AI facilitator. The AI could prompt quiet participants, interrupt overly dominant voices, display real-time transcripts, and guide the group toward consensus. Once proposals were agreed upon, they were synthesized and voted on in a plenary session.
Several notable policy ideas emerged from the discussions: one group proposed requiring digital signatures for all advertisements to verify authenticity; another advocated holding platforms like Facebook financially liable for damages caused by scam ads; a third group suggested throttling data speeds for companies that refused to comply with regulations.
These ideas gained overwhelming public support—more than 85% approval—and within months, legislation was drafted, passed, and implemented. By July, the new regulations were in force, and major platforms took steps to prevent further abuse. Tang noted that since the law's enactment, scam ads had virtually disappeared from Taiwanese social media.
Tang framed this example as a model for how AI can support democracy by helping large groups deliberate effectively, reach consensus, and rapidly translate public opinion into policy. She emphasized that the AI tools did not replace human judgment but augmented it—allowing citizens to collaborate at scale without descending into chaos or polarization.
The session concluded with a broader reflection on how similar systems, including citizen assemblies and future design councils, are already familiar in Japan. With the help of AI, Tang argued, such deliberative frameworks could be made even more inclusive, effective, and responsive.
Source: テレ東BIZ
