EDDIE YANG (PURDUE UNIVERSITY)

December 18, 2025 at 1:15 PM UTC

Online via Webex. Please register here:fu-berlin.webex.com/webappng/sites/fu-berlin/webinar/webinarSeries/register/4a9b46cc059949ec85b7360d963cca0a There has been a flurry of recent concern about the question of who directly controls large language models. We show through six studies that coordinated propaganda from powerful global political institutions already indirectly influences the output of U.S. large language models (LLMs) via their training data, a pattern which is easiest to see in China. First, we demonstrate that material originating from China's Publicity Department appears in large quantities in open-source pre-training datasets. Second, we connect this to U.S.-based commercial LLMs by showing that they have memorized sequences of propaganda, suggesting that it does appear in their training data. Third, we use an open-weight LLM to show that additional pre-training on Chinese state propaganda generates more positive answers to prompts about Chinese political institutions and leaders---evidence that propaganda itself, not mere differences in culture and language, can be a causal factor in the behavioral differences we observe across languages. Fourth, we show that prompting commercial models in Chinese generates more positive responses about China's institutions and leaders than the same queries in English. Fifth, we show that this language difference holds in prompts of actual Chinese-speaking users. Sixth, we extend our findings with a cross-national study that indicates that the languages of countries with lower media freedom show a stronger pro-regime valence than those with higher media freedom. Finally, we show results that demonstrate that the phenomenon described here is broader than propaganda and state media alone. Our findings join the ample recent work demonstrating the persuasive power of LLMs. Together, these results suggest the troubling conclusion that states and powerful institutions will have increased strategic incentives to disseminate propaganda in the hopes of poisoning LLM training data. Bio: Eddie Yangis an Assistant Professor of Political Science and faculty member in the Institute for Physical Artificial Intelligence at Purdue University. He received his Ph.D. in political science from the University of California San Diego. Yang studies the politics of innovation and technology. His research has been published at theProceedings of the National Academy of Sciences, andPolitical Analysis, among other outlets.

主讲人： 唐晓玮|XIAOWEI TANG 哈佛大学文理学院访问博士，中央美术学院博士生 黄晟昱|DR. SHENGYU HUANG 哈佛大学文理学院博士后研究员，NOLIBOX.AI联合创始人 主持人/评议人： 范敏 教授|PROF. MIN FAN 麻省理工学院媒体实验室客座教授，中国传媒大学动画与数字艺术学院教授 张琴|DR. QIN ZHANG 哈佛大学CAMLAB合作研究员，哈佛大学费正清中心合作研究员

December 19, 2025 at 8:00 PM UTC

本讲座将从脑机接口的前沿创意出发，探讨人类感知、认知与审美之间的关系，结合人工智能的发展，解析“审美智能”的形成机制与未来应用前景。本次讲座为线下讲座。

JEFFREY DING (GEORGE WASHINGTON UNIVERSITY)

January 15, 2026 at 1:15 PM UTC

Online via Webex. Please register here:fu-berlin.webex.com/webappng/sites/fu-berlin/webinar/webinarSeries/register/4a9b46cc059949ec85b7360d963cca0a Emerging economies face significant challenges in managing safety risks from powerful technological systems. Indeed, many analysts have identified China as the most likely source of a major accident linked to emerging technologies. Yet, contrary to these expectations, China has achieved a remarkable safety record in certain technological domains, such as civil aviation and nuclear power. How? We theorize that, for industries in which one firm’s accident damages the reputation of all others, international industry associations can contribute to improved safety standards in emerging economies. When firms share a collective reputation, industry associations exert positive peer pressure by subsidizing laggards’ efforts to raise their safety standards and protecting members from public naming and shaming. This departs from existing theories of international private regulation on certification clubs that set strict quality, safety, and environmental standards to deny association benefits to non-members. To demonstrate differences between these two mechanisms, we examine interactions between international industry associations and Chinese firms in three high-risk technological domains: nuclear power, civil aviation, and chemicals. Our findings have implications for scholars interested in the interdependencies between international public regulation and private regulation as well as policymakers trying to manage the safety risks of emerging technologies such as artificial intelligence. Bio: Jeffrey Dingis an Assistant Professor of Political Science at George Washington University. Previously, he was a postdoctoral fellow at Stanford's Center for International Security and Cooperation, sponsored by Stanford's Institute for Human-Centered Artificial Intelligence. He researches great power competition and cooperation in emerging technologies, the political economy of innovation, and China's scientific and technological capabilities. His book,Technology and the Rise of Great Powers(Princeton University Press, 2024), investigates how past technological revolutions influenced the rise and fall of great powers, with implications for U.S.-China competition in emerging technologies like AI. Other work has been published or is forthcoming inEuropean Journal of International Relations,European Journal of International Security,Foreign Affairs, International Studies Quarterly,Review of International Political Economy, andSecurity Studies, and his research has been cited inThe Washington Post,The Financial Times, and other outlets.

DAVID Y. YANG (HARVARD UNIVERSITY)

January 22, 2026 at 1:15 PM UTC

Online via Webex. Please register here:fu-berlin.webex.com/webappng/sites/fu-berlin/webinar/webinarSeries/register/4a9b46cc059949ec85b7360d963cca0a Venture capital plays an important role in funding and shaping innovation outcomes, characterized by investors’ deep knowledge of the technology, industry, and institutions, as well as their long-running relationships with the entrepreneurship and innovation community. China, in its pursuit of global leadership in AI innovation and technology, has set up government venture capital funds so that both national and local governments act as venture capitalists. These government-led venture capital funds combine features of private venture capital with traditional government innovation policies. In this paper, we collect comprehensive data on China’s government and private venture capital funds. We draw three important contrasts between government and private VC funds: (i) government funds are spatially more dispersed than private funds; (ii) government funds invest in firms with weaker ex-ante performance signals but these firms exhibit growth rates exceeding those of firms in which private funds invest; and (iii) private VC funds follow government VC investments, especially when hometown government funds directly invest on firms with weaker ex-ante performance signals. We interpret these patterns in light of VC funds’ traditional role overcoming information frictions and China’s unique institutional environment, which includes important frictions on mobility and information. Bio: David Y. Yangis a Professor in the Department of Economics at Harvard University and Director of the Center for History and Economics at Harvard. David is a Faculty Research Fellow at NBER, a Global Scholar at CIFAR, and a fellow at BREAD. David’s research focuses on political economy. In particular, David studies the forces of stability and forces of changes in authoritarian regimes, drawing lessons from historical and contemporary China. David received a B.A. in Statistics and B.S. in Business Administration from University of California at Berkeley, and PhD in Economics from Stanford.

NICHOLAS DE VILLIERS

January 22, 2026 at 11:30 PM UTC

In Cruisy, Sleepy, Melancholy: Sexual Disorientation in the Films of Tsai Ming-liang(University of Minnesota Press, 2022), Nicholas de Villiers contends that we need to theorize both queer time and space to understand Taiwan-based director Tsai Ming-liang's cinematic explorations of feeling melancholy, cruisy, and sleepy.Building on those arguments, this presentation starts with a reading of Tsai’s short filmIt’s a Dream(2007)—set in a movie theater in Malaysia—as a microcosm of Tsai’s themes and motifs of sleep/dreaming, cruising, nostalgia, and the space of the cinema. It then addressesTsai’s “post-retirement” (after 2013) filmsand museum installations, including the queer Teddy award-winning digital feature filmDays(Rizi, 2020) shot in Taiwan, Hong Kong, and Thailand, and the short filmThe Night(2021) shot in Hong Kong in 2019. Both were featured in the solo exhibitionTsai Ming-liang’s Daysat the Museum of National Taipei University of Education (MoNTUE) in 2023, experimenting with "expanded cinema" and collaborative curation of Tsai's still expanding body of work. Nicholas de Villiersis Professor of English and film at the University of North Florida and 2023–2024 Fulbright U.S. Senior Scholar in Taiwan at National Central University at the Center for the Study of Sexualities. He is the author ofOpacity and the Closet: Queer Tactics in Foucault, Barthes, and Warhol(2012),Sexography: Sex Work in Documentary(2017), andCruisy, Sleepy, Melancholy: Sexual Disorientation in the Films of Tsai Ming-liang(2022), all from the University of Minnesota Press.