Without adequate defenses, AI technology could undermine social trust and democratic institutions, making it critically important to strengthen whole-of-society defense resilience. That came from Vice President Hsiao Bi-khim (蕭美琴) at the opening ceremony of this year’s Taiwan AI Academy Annual Conference on Tuesday, August 18.

Hsiao said that AI sovereignty cannot be built on hardware advantages alone, but must also include data governance and cultural values. Towards that goal, the government is advancing the Ten Major AI Infrastructure Projects, which focus on computing power, data governance, talent cultivation, and innovation ecosystems, as well as future-looking technologies like quantum computing, silicon photonics, and intelligent robotics. Hsiao added that the Cabinet is actively cataloging corpora and datasets that reflect Taiwan’s values for use in training sovereign AI. In addition, it is also improving data governance and supporting open data initiatives in business, technology, humanities, public administration and healthcare.

Hsiao said that AI brings both convenience and challenges, like disinformation, deepfake technology, cognitive warfare, and cyberattacks. She said: "If a country lacks the capacity to prevent AI from generating massive amounts of disinformation and deepfakes used in cognitive warfare, or cyberattacks targeting infrastructure, then even the most advanced technology can become a double-edged sword eroding social trust and democratic systems. This is why the president has emphasized whole-of-society resilience."

Hsiao said the government has incorporated information, communications, and infrastructure resilience into its whole-of-society defense resilience framework. It is also advancing trustworthy AI development that is human-centered and protects human rights, balancing innovation and security under the Artificial Intelligence Basic Act.