Anthropic Visits National Chengchi University for Interdisciplinary Talks on AI Governance and Democratic Resilience
【Article by OIC】
A delegation from leading artificial intelligence company Anthropic, led by Chief Global Affairs Officer Dr. Tino Cuéllar, made its first visit to National Chengchi University (NCCU) on November 13. NCCU President Tsai-Yen Li, joined by five student representatives, welcomed the delegation alongside faculty from the Office of International Cooperation, the College of Communication, the Department of Economics, the College of Law, and the College of International Affairs. The two sides held in-depth discussions on recursive self-improvement, AI safety governance, electoral integrity, democratic resilience, and geopolitical and labor market impacts, reaching a strong consensus on future academic research and election information security collaboration in Taiwan.
In his welcoming remarks, President Tsai-Yen Li noted that as NCCU approaches its centennial next year, its long-standing strength in the humanities and social sciences—which has produced countless leaders across government and high-tech sectors—is more vital than ever. In the AI era, he emphasized, these disciplines not only drive technological adoption, but also determine how society ultimately puts AI to use. Closing the meeting, President Li compared AI to an automobile: "AI technology is like a car's engine, and humans must remain the drivers. We decide how to steer and when to press the gas or the brakes. This is NCCU’s core mission in cultivating humanities and social science leaders: keeping humans in the loop so that technology brings societal well-being rather than disaster."
Dr. Tino Cuéllar explained that Anthropic was founded on the philosophy of bringing together engineers and physicists alongside voices from journalism, literature, and the humanities to ensure AI scales safely through multidisciplinary perspectives. Highlighting three major trends shaping the field, he noted that AI has entered a phase of "recursive self-improvement"—with Anthropic’s model Claude already assisting in roughly 25% of internal R&D and up to 80% of coding—making strict human oversight essential. Beyond accelerating biomedical breakthroughs and bridging gaps in rural healthcare and personalized education, he stressed that global governance requires clear accountability and legal liability, positioning Taiwan—with its semiconductor ecosystem and deep technical expertise—as an indispensable partner.
During the meeting, Professor Shinn-Shyr Wang, Chair of NCCU’s Department of Economics, gave a presentation titled "One Chip, Four Questions." Building on Taiwan’s pivotal role in manufacturing over 90% of the world’s advanced AI chips, he outlined four collaborative initiatives. First, he proposed drawing on more than three decades of polling expertise from NCCU’s Election Study Center to integrate official voting prompts into AI models during elections and publish Traditional Chinese post-election interference reports. Second, echoing Taiwan's Artificial Intelligence Basic Act, he suggested pairing NCCU’s economics and legal scholarship with local industry data to measure AI’s impact on entry-level jobs. Third, he proposed leveraging over 70 years of think-tank experience at NCCU’s Institute of International Relations to co-develop risk models for the Taiwan Strait and semiconductor supply chains, alongside a fourth initiative exploring real options strategies for corporate AI investments. He also invited Anthropic to join the 2027 international conference on "AI and Work," commemorating NCCU’s centennial and the Department of Economics' 60th anniversary.
In the open exchange, Distinguished Professor Yi-Ning Katherine Chen of the College of Communication—who also serves on Meta’s Oversight Board—drew on her experience in global platform governance to examine how AI models navigate political criticism and content boundaries across different linguistic and political contexts. Professor Cho-Hsin Su of the College of International Affairs followed with questions on tech diplomacy amid great-power competition and supply chain interdependence. In response, the Anthropic team noted that Taiwan’s deep experience in cybersecurity and content policy offers valuable lessons, reaffirming its commitment to strengthening model performance in Traditional Chinese while upholding democratic values.
NCCU's student representatives also brought sharp interdisciplinary perspectives to the table. A student from the Department of Chinese Literature asked how "Constitutional AI," rooted in a Western framework, could incorporate Eastern cultural perspectives, while a law student examined service stability and how democratic allies can avoid vendor lock-in. Meanwhile, a representative from the International College of Innovation focused on the long-term effects of AI agents on public deliberation and democratic institutions, and a Management Information Systems student inquired how corporate internal safety policies can go beyond existing regulations to enforce self-restraint.
Dr. Cuéllar commended the students for the depth of their inquiry. "The technological responsibilities of this generation will soon be passed into your hands," he said. "Seeing NCCU students thinking about the right—and most critical—questions gives me great confidence in the future." The meeting concluded warmly, with both sides agreeing to maintain close contact through designated liaisons to advance concrete transnational and cross-disciplinary collaboration.