Events

The Silicon Gaze: A typology of biases and inequality in LLMs through the lens of place
This paper introduces the concept of the silicon gaze to explain how large language models (LLMs) reproduce and amplify long-standing spatial inequalities. Drawing on a 20.3-million-query audit of ChatGPT, we map systematic biases in the model's representations of countries, states, cities and neighbourhoods. From these empirics, we argue that bias is not a correctable anomaly but an intrinsic feature of generative AI, rooted in historically uneven data ecologies and design choices. Building on a power-aware, relational approach, we develop a five-part typology of bias (availability, pattern, averaging, trope and proxy) that accounts for the complex ways in which LLMs privilege certain places while rendering others invisible.
Analysing the US-China "AI Cold War" Narrative
Discussions about artificial intelligence (AI) are gaining prominence in the recent revival of “cold war” narratives comparing US-China relations today to the historical rivalry between the US and the Soviet Union. Drawing on a review of existing academic and policy literature engaging with the “AI cold war” narrative, this paper examines how the narrative is justified, and numerous ways that it can be challenged. It finds that the framing is largely driven by the securitisation of AI, as state actors and policy pundits view AI innovations' dual-use capabilities as key to national security and ideological competition. However, critics posit that the narrative exaggerates China's AI capabilities, promotes commercial interests of tech firms and defence contractors, creates self-reinforced militarisation, and undermines the potential for international research and regulatory cooperation. Moreover, the Cold War binary framing may misrepresent the global distribution of AI capabilities. To extend beyond the AI cold war narrative, future research may recognise the limitations of the binary framing and expand analysis on the AI development strategies of third-party players (including those from the Global South) drawing upon local and regional political economic dynamics and development contexts. This paper concludes by inviting scholars to rethink the affective power of narratives and contribute research and narrative analysis that allow for the articulation of perspectives from third countries.

Corruption is often treated as a side-topic in public policy debates, yet in practice it is the line between effective policies that flourish and ambitious reforms that fail. Whether your focus is climate policy, social welfare, urban development, digital governance, or any other policy domain, this seminar will equip you with an essential skill set: recognizing the hidden links between policies and corrupt incentives, understanding the range of corruption-resistant interventions that countries around the world are experimenting with, and gaining insights into the political and institutional realities that shape implementation.
From lobbying and influence practices to outright corruption and money laundering, the seminar will draw on concrete cases, introduce new ways to measure integrity risks, and explore some of the most innovative tools, such as data analytics, corruption risk management, business integrity, and more. Expect a highly applied session that confronts the sometimes uncomfortable truth about what works, what doesn’t, and why.
Join us to discover why integrity competence is not just an academic concern, it is a strategic advantage for every public policy practitioner

The seminar examines the October 2021 announcement of Hong Kong’s Northern Metropolis Plan as a quasinatural experiment to assess short-run impacts on household behavior. Using difference-in-differences on 100,576 housing transactions, 4.7 million consumption records, and demographic indicators across 18 districts, Prof. Fan documents a 3.9% increase in housing prices in the Northern Metropolis one year after the announcement, with a sharper 6.9% rise in the first four months. Over the same initial period, local consumption increased by 4.1%, consistent with anticipated housing-wealth effects and concentrated in the private housing market. Demographically, the treated region saw inflows of higher-educated, higher-income households, more families with children, and rising school enrollments; there are also spillovers to Shenzhen’s housing market. While cross-region gaps appear to narrow in the short term, within-region inequality may widen, as early gains skew toward higher socioeconomic groups, underscoring policy needs around affordability and welfare.

The rapid ascent of artificial intelligence, big data, and cloud computing is fundamentally reshaping societies and economies worldwide. For the Chinese government, which has actively championed a state-led model of digital transformation, these technologies present a dual-edged sword: unprecedented tools for enhancing governance capability and significant challenges to established bureaucratic structures and policy paradigms.
 
This seminar examines the critical intersection of digital innovation and public administration in China. It explores the central question: How can the Chinese state cultivate "agile governance" to harness the benefits of intelligent technologies while mitigating the associated risks? We will first analyze the progress made through initiatives like "Internet + Government Services" and the "Social Credit System," which aim to create a more responsive, data-driven, and efficient state.

 

Amid escalating geopolitical tensions, international collaboration and researcher mobility between the United States and China have become both crucial and constrained. This talk presents a series of empirical studies examining how flight connectivity, talent flows, and judicial verdicts intersect to shape the dynamics of Sino-US scientific exchange.

Large language models (LLMs) are increasingly integrated into judicial processes, assisting with research, summarization, and drafting. However, their unverified use threatens core principles, including accountability, transparency, and ultimately public trust in the judiciary. Our study demonstrates interpretive divergences across leading LLMs (GPT‑4, Claude, Gemini, LLaMA) on benchmarked legal tasks, revealing systematic vulnerabilities that we term omission, injection, and framing. Through controlled perturbations, semantic analyses, and cross-model evaluations on CaseHOLD and LexGLUE, we quantify semantic drift and show that even small variations in prompts or model design can materially alter legal reasoning. To address these risks, we situate our research within broader scholarship in law and the social
sciences on judicial behavior. Building on this foundation, we propose a governance framework designed to ensure that LLM adoption strengthens, rather than compromises, the fairness and legitimacy of judicial decision-making, thereby safeguarding public trust in the judiciary. This paper was presented at one of the workshops held in conjunction with IJCAI 2025 in Montreal, where it received a Best Paper Award and is forthcoming in a special issue of Springer Lecture Notes in Artificial Intelligence (LNAI).

The climate crisis calls for a massive and rapid retooling of our economy and society. Using neo-Schumpeterian history of technological revolutions and organizational theory, I identify four alternative pathways and assess their relative merits and tradeoffs. This suggests opportunities for collaboration and mutual learning across alternative systems of societal governance.

Communities today face intertwined risks from climate change, public health crises, and rapid social transformation, demanding more adaptive governance strategies. This presentation examines how integrating artificial intelligence with human expertise can strengthen resilience in practice. Drawing on experiments in major Chinese cities, it highlights how AI tools—when aligned with human judgment—enhance risk anticipation, resource allocation, and participatory decision-making. Rather than replacing human agency, AI serves as a complement, fostering institutional adaptability and public trust. The discussion emphasizes practical lessons for policymakers and urban leaders seeking to build robust, inclusive, and future-ready communities in China and beyond.

The growing aging population has drawn increasing attention to aging arrangements and mental health issues among older adults. However, despite the rapid growth of ethnic minority (EM) older populations alongside overall demographic aging, their distinct needs remain largely neglected in policy planning and service provision. While extending literature has focused on discrimination, poverty, and ethnic inequality, this study adopts a more comprehensive, culturally informed perspective to examine EM aging and mental health challenges, and to propose policy-level strategies to address these issues.

To have a comprehensive view of EM aging and mental health challenges, this thesis builds on an ecological framework to enable a multi-level analysis, spanning individual, interpersonal, community, and policy dimensions. Through interviewing EM older adults in Hong Kong, the first study investigates the factors influencing their preferences for aging-in-place arrangements. The second study takes a global perspective, identifying the culturally shaped coping strategies they employed in response to mental health challenges. Recognizing the importance of traditional family values among ethnic minorities, the third study using self-collected survey data examines the mechanisms linking public transfers, private financial support, and the mental well-being of EM older adults.

Collectively, the findings highlight the critical importance of cultural inclusiveness in aging policy making. This thesis contributes not only empirical evidence on aging and mental well-being among ethnic minorities, but also emphasizes the crucial role of cultural context in shaping EM mental health, influencing decisions related to aging arrangements, and determining the effectiveness of social protection policies.

The expansion of data centers and telecommunication networks has raised environmental concerns driven by surging electricity consumption and corresponding greenhouse gas (GHG) emissions. While the environmental impacts of ICT infrastructure are acknowledged, monitoring emissions and policy design remain underdeveloped. This dissertation examines public policy challenges in cloud emission estimation addressing three parameters: (1) existing policy measures to mitigate data center impacts, (2) a novel carbon accounting framework for cloud emissions (consumption-based), and (3) implementation pathways for this framework.
Parameter (1) is assessed via qualitative content analysis of six data center hubs: Singapore, the Netherlands, Ireland, Germany, the USA, and the UK. The study categorizes policy responses as disruptive (halting new licenses), adjustment (implementing stricter standards), or continuation (maintaining deployment despite environmental impacts). The analysis indicates that stricter energy accuracy standards support growth planning, while policies align infrastructure strategies with available electricity capacity.
For parameter (2), the research examines the limitations of the traditional, location-based carbon allocation method. The method overlooks the geographical separation between emission sources and cloud consumers. Using the consumption approach for three of the six hubs (Germany, Ireland, and the Netherlands) from 2017 to 2022 reveals differences in consumption profiles: Germany’s cloud emissions are 75% domestic, Ireland’s 85% foreign, and the Netherlands’ split evenly, highlighting the need for equitable emission accountability.

To address parameter (3), the study identifies four estimation risk sources: emission factors, methods, data sources, and boundary definitions. These risk sources stem from the location-based method that overlooks the geographical separation between cloud hosts and consumers, which risks disproportionately allocating emissions only to cloud-hosting jurisdictions. The study proposes three policy interventions to mitigate the risks: stricter accounting rules, eco-labeling, and carbon border adjustments. The findings of the three dimensions are summarized into a decision framework to guide governments in monitoring and mitigating cloud emissions.

Singapore's primary care system faces challenges related to fragmentation and misaligned incentives despite impressive health outcomes. This paper examines the system through four dimensions: governance, provision, financing, and payment. It argues that insufficient coordination and limited government influence over private providers hinder efforts to strengthen primary care. Policy interventions focused on aligning incentives and strengthening governance are crucial for building an effective primary care system.

This paper quantifies the effect of distributed solar generation on distribution feeders using unique, proprietary data from individual solar installations and feeders in Connecticut. We find that each additional kilowatt (kW) of distributed solar reduces the annual maximum feeder load and the top one percentile of feeder load by 0.11 kW. Using a matched difference-in-differences approach, we also find evidence of a 3.6% solar rebound effect, primarily occurring in spring and fall. Importantly, this rebound does not affect the impact of distributed solar on peak feeder loads, as these predominantly occur during the summer. Except in high-congested feeders, the economic value of deferred distribution capacity investment ranges from $0.3 to $2.9 per MWh, significantly below the cost premium of distributed solar compared to utility scale solar. Our findings imply that even in a summer-peaking system, distributed solar generation cannot be rationalized solely by savings from deferred distribution feeder investments. 

The two universities cooperate again to host The Policy Dialogue Series which began in 2018-19 and is now in its seventh annual series. Professor Anthony Cheung and Professor Lui Tai-lok will be joined by other speakers in the relevant fields to explore some critical issues and public policy challenges facing Hong Kong. The Policy Dialogue Series is interactive and encourages the participation of the HKUST and EdUHK communities as well as members of the public who are concerned about Hong Kong’s future.

Geopolitics is redefining the world that Hong Kong used to face and operate in. At the same time, the city is further integrated into the national development trajectory, especially within the Guangdong-Hong Kong-Macao Greater Bay Area (GBA), which opens new opportunities as well as complementarities and competitions vis-à-vis other Chinese cities. Also, both government and business now seek to strengthen economic links with Southeast Asia and the Middle East. How should Asia and GBA feature in Hong Kong’s rethinking of regionalization strategy to drive new growth?

Texts such as news articles, social media posts, contracts, financial reports, laws, and regulations are deeply embedded in the real-world dynamics of the economy and business. Traditionally, economic studies have had limited means of systematically and quantitatively analyzing such textual data, primarily through text mining and certain qualitative methods, making it difficult to conduct complex, context-aware analysis. However, generative AI, including large language models (LLMs), has the potential to dramatically transform this landscape. This talk explores how generative AI can contribute to policy analysis, with a focus on applications in economic policy studies, including monetary and fiscal policy analysis.

In the last decades of the past century, growing criticism of the effects of government intervention in the economy led to a decline in the governments' role in the production of goods and services. In the area of local government and public service delivery, a wave of privatization was the main tool of reform, often through the contracting out of public service delivery. Because the results of privatization were much more modest than expected, other types of reforms have intensified in the first decades of this century. On the one hand, regarding the jurisdiction of the provider (amalgamation and intermunicipal cooperation); on the other, regarding the organization of production (corporatization and remunicipalization). After five decades of local government reform, a local machine-service hybridization is the most common approach to public service delivery.

The world has experienced different moments in great power politics. Such as bipolarity (1945-1989), unipolarity (1989-2014) and we have entered a multipolar phase (2014-onwards). 
 
In this phase, the United States of America, the People´s Republic of China and the Russian Federation are the three most relevant actors that seek to maintain and increase their international power. 
 
This conference focuses on the current moment, starting by the second term for President Donald Trump which is speeding up international change. The weakening of the transatlantic alliance, following the wars in Ukraine and Palestine takes the international environment to a different scenario which is worth analyzing.
 
Which includes President Trump´s expansionist intentions with Greenland, Canada and the Panama Canal, as well as the European Union´s intention to form a stronger alliance that will be most likely stopped by the competition between Germany and France. 
 
These topics will be developed using Theory of International Relations. 

 

In his major address to Hong Kong on 1 July 2022, President Xi Jinping emphasized the need to care about the young generation and expressed hope that all of Hong Kong’s young people “will devote themselves to building Hong Kong into a better home, writing a rewarding chapter of their life with impassioned youth” (投身到建設美好香港的行列中來，用火熱的青春書寫精彩的人生). Young people are the future of our society. What are their aspirations and most pressing concerns today? How can a more conducive environment be fostered to enable them to build their hope and stake in the city?

We use large language models to rank how advanced China, the US and the EU are in key critical technologies by using patent data but control for the relevance of each patent. We also look into the speed of spillovers and which companies dominate for each country. This analysis is followed by policy conclusions for the three economic areas analyzed.

We are thrilled to invite you to an engaging career talk titled "The Career Playbook: Tips and Insights from Policy Professionals," featuring three dynamic mid-career policy professionals from leading think tanks and major corporations. This is a unique opportunity to connect with industry leaders right here at the Clear Water Bay campus and gain invaluable insights into the world of public policy.

Join us for an inspiring session where our esteemed speakers will share their personal journeys, offering practical tips on how to effectively balance study and work, hone your internship skills, navigate the job-hunting landscape, and strategically plan your career development. They will also delve into their experiences conducting real-world policy analysis, providing you with a behind-the-scenes look at the challenges and triumphs faced in the field.

Don’t miss out on this chance to learn from the best and equip yourself with the knowledge and skills needed to excel in your future career. This talk promises to be both informative and motivating, so seize the opportunity to enhance your professional journey and network with accomplished professionals. We look forward to seeing you there!

There will be lots of time for questions and discussion. In this seminar, two distinguished female scholars, both of whom have held ministerial positions in South Korea and Hong Kong SAR, will share their unique experiences of navigating the intersection between academia and government.
The two Sophias will also engage in a discussion on critical social and health policy issues facing these two Asian tiger economies.

Hong Kong faces various demographic challenges including an ageing society, decline in birth rate, middle-class brain drain, and talent and labour shortage. As early as 2002, a government task force recommended setting population policy targets to respond to changing demographic conditions and market situations. How has Hong Kong’s population policy fared over the past two decades — to address demographic issues, reorganise and improve services, and groom and attract a broad range of talent necessary for social and economic development?

Mark Jit will briefly discuss the current state of the field of global health in today’s changing and uncertain world. Prof. Jit will then talk about some of the challenges and opportunities facing an academic global health department, using NYU’s recently established department as an example. There will be lots of time for questions and discussion.