↓ 跳过正文

AI相关

The Safety Tether: How China Manages Civil Society Through Lawfare

GOV 39/3 2026-06-19 Original The Safety Tether: How China Manages Civil Society Through Lawfare Timothy Hildebrandt, Blake Miller Abstract ABSTRACT How do authoritarian regimes benefit from civil society organizations' expertise and capacity while mitigating the potential risks they pose? Using machine learning and interrupted time series modeling of an original custom‐annotated dataset of all historical NGO registrations in China, we find that NGO policy is unevenly implemented. This unevenness, however, is not random. China uses NGO registration policies to facilitate local risk utility tradeoffs in line with threat perceptions and governance needs. This, we argue, is a form of lawfare—using legal means to achieve political or repressive ends. Examining a 2013 policy shift, we find local governments relax sponsorship requirements more in policy domains with low state capacity and less where civil society poses a threat, tailoring national mandates to shape local conditions. We show how lawfare is a complex instrument providing local flexibility in repression and cooptation of civil society, under the guise of routine legal and administrative procedures.

AI Public Value Creation: Data Encoding, Aggregation, and Algorithmic Computation

PA — 2026-06-08 Original AI Public Value Creation: Data Encoding, Aggregation, and Algorithmic Computation Antonio Cordella, Francesco Gualdi Abstract ABSTRACT AI systems in public administration challenge the premise that value creation arises from managers' strategic mediation among the vertices of Moore's strategic triangle: Public value, legitimacy, and operational capacity. AI imposes new logics that reshape vertices and their interaction dynamics. Critically, these transformations become entrenched through what we term algorithmic public value lock‐in: the crystallization of contingent value definitions into durable technical infrastructures resistant to democratic revision. Examining Peru's SISFOH, an algorithmic welfare‐targeting system, we reveal three interconnected shifts: data management recasts legitimacy as computational objectivity; standardized classifications redefine public value in narrow, quantifiable terms; and algorithmic processing transforms operational capacity from adaptive discretion to rigid execution. Together, these shifts produce lock‐in effects that potentially insulate public value creation from political negotiation.

Negativity trumps source: exploring the perceptions of AI-advice among public officials

PMR — pp. 1-28 2026-06-07 Original Negativity trumps source: exploring the perceptions of AI-advice among public officials Yuyan Cao, Xin Liang Abstract The advancement of artificial intelligence (AI) may renew our understanding of evidence-based policymaking. Grounded in motivated reasoning and negativity bias literature, this study adopts a 2 × 2 survey experiment to explore how advice source and valence affect public officials’ trust and policy preference. Findings demonstrate that advice containing negative result exerts disproportionate influence. Moreover, when advice aligns with public officials’ prior policy beliefs, negativity bias amplifies their preference. These findings highlight how cognitive biases impede the public officials’ interpretation to policy advice.

Framing AI in public innovation: the engagement-manipulation nexus

PMR — pp. 1-25 2026-04-16 Original Framing AI in public innovation: the engagement-manipulation nexus Adedapo Adebajo, Ngozi Okpara, Aisha Kolawole Abstract AI integration into public service innovation creates a paradox: promising enhanced engagement while risking manipulation. This paper develops the Engagement-Manipulation Nexus, a conceptual framework using Framing Theory to analyse this tension. We adapt Framing Theory to the algorithmic era, arguing AI is both an object to be framed and an agent disseminating frames. AI innovation trajectories are forged through communicative struggles among actors deploying competing frames. This paper contributes a novel theoretical framework, theorizes public innovation as communicatively constructed, and provides a diagnostic tool for practitioners. Critical frame awareness is essential for steering AI towards public value and democratic governance.

Smiley bots, satisfied citizens? The impact of AI humanization on citizen experience in public services

PMR — pp. 1-31 2026-04-02 Original Smiley bots, satisfied citizens? The impact of AI humanization on citizen experience in public services Jinjin Wu, Yifan Chen Abstract This study examines how AI humanization influences citizen experience in public service delivery by integrating the Stereotype Content Model (SCM) and Emotion as Social Information (EASI) theory. Focusing on AI chatbots as a research context, we explore how humanized responses shape perceived warmth and competence and whether effects differ between programmed and non-programmed services through a survey experiment with four vignette-based scenarios. Findings underscore the importance of context-sensitive and carefully calibrated AI design and functioning choices. By linking SCM’s content dimensions with EASI’s affective and inferential mechanisms, the study provides a social-psychological lens to understand citizen–government interactions in AI-driven public administration and provides practical guidance for designing AI-enabled services that remain citizen-friendly while safeguarding core public values.