Collaborative research programs dedicated to investigate the relations between technology and civic engagement and other challenges around concepts like citizenship, participation, and democracy.
Researchers developed a value-based method for prompting LLMs that uses culturally grounded survey-derived value profiles instead of demographic proxies, significantly improving the models’ ability to predict and reproduce diverse population opinions across countries, especially in underrepresented cultures.
This paper argues that accounts of deliberative mini-publics misdiagnose the crisis by framing it primarily as a procedural deficit, neglecting a deeper dimension that populism exploits: the erosion of collective political agency.
We argue that the use of LLMs as proxies for participants enacts the surrogate effect, raising ethical and epistemological concerns that extend beyond the technical limitations of current models to the core of whether LLMs fit within qualitative ways of knowing.
Can a Structured Governance Memory System (GMS) be Designed and Implemented to Enable Validator-Governed Protocols and DAOs to Make Better-Aligned Decisions that Avoid Unintended Consequences, Preserve Institutional Learning, and Increase Long-Term Resilience and Adaptability?
Using global case studies from the Participedia database, the findings reveal that democratic innovations excel in various attributes but rarely deliver substantial redistributive impact.
Our findings suggest that the persuasive power of current and near-future AI is likely to stem less from model scale or personalization and more from post-training and prompting techniques that mobilize an LLM’s ability to rapidly generate information during conversation.
These findings support preference- and incentive-based accounts of participation but suggest that light-touch interventions are unlikely to bridge participation gaps, let alone polarization.
The Climate Citizens’ Assembly Kawasaki (CCAK) was implemented in mid-2021 to produce recommendations for the local government to incorporate into its basic climate plan.
As journalists look for different ways to tell stories and rebuild communities, this paper details an action research approach to a unique live journalism event in a public space.
"[T]his article proposes an open government framework that integrates multiple concepts related to open government and categorizes them as either constitutive components or potential results"
Comparing traditionally structured survey questions vs. the benefits of open-ended survey questions combined with text analysis to allow a wider range of responses.
AI technologies can empower public agencies by enabling real-time network monitoring, predictive analytics for risk prevention, and automated responses to cyber incidents
This is a regularly updated collection of academic studies and industry reports about digital deception. It currently includes short descriptions of 55 academic studies and systematic reports.
We provide initial evidence that people’s ability to spot and resist misinformation improves after gameplay, irrespective of education, age, political ideology, and cognitive style.
We find that Go Viral!, a novel five-minute browser game, (a) increases the perceived manipulativeness of misinformation about COVID-19, (b) improves people’s attitudinal certainty (confidence) in their ability to spot misinformation and (c) reduces self-reported willingness to share misinformation with others.
This study examines 51 Freedom of Information requests to reveal how the UK’s Department for Work and Pensions uses opaque data-driven fraud detection systems in welfare, highlighting the limited transparency, the role of journalists and charities in uncovering details, and inconsistencies in the department’s disclosures.
This paper argues that automated decision-making in UK public administration lacks adequate scrutiny, and proposes regulatory safeguards through mandatory pre-deployment impact assessments and algorithmic auditing to strengthen accountability in Parliament and the courts.
"By conducting a national-scale field experiment among 949 provincial-level agencies in China....results show that legal regulative pressure and social normative pressure make agencies more likely to respond within the legal timeframe and provide the requested information."
This paper examines emerging experiences with chatbots in government interactions, with a focus on exploring what public administration practitioners and scholars should expect from chatbots in public encounters.
Here, we explore four applications of LLMs to improve digital public squares: collective dialogue systems, bridging systems, community moderation, and proof-of-humanity systems.
Analyzes the size, rhetorical nature, and non-violent nature of the pro-Palestine protest wave, which it finds to be "the largest, most sustained US protests sparked by a foreign event" since their data became available in 2017. Published in Social Movement Studies
We propose an alternative platform model that the social fabric an explicit output as well as input. By E. Glen Weyl, Luke Thorburn, Emillie de Keulenaar, Jacob Mchangama, Divya Siddarth, Audrey Tang
In this paper, building on insights from the theory and practice of deliberative democracy, we provide a "Democracy Levels" framework for evaluating the degree to which decisions in a given domain are made democratically.
This article proposes, and presents results from, an online technology platform and a methodology, which can help depolarize sectors of society, with the specific application to the current divisive post-war landscape within Colombia.
Google DeepMind and Stanford researchers built the "Habermas Machine" and found that AI mediators generated more palatable summary statements of the discussions, as rated by participants, than human-written summaries, while still representing minority views in the final version.
The 2023 edition of the “IPIC – Portuguese Municipal Councils’ Online Presence Index” series of studies presents the results of the evaluation of Portuguese municipal councils’ websites.
Waymo partnered with dashcam company Nexar to analyze 500 million miles of driving, including 335 crashes, to create the 'largest ever’ dataset of pedestrian and cyclist injuries
"Importantly, people experiencing distress are more likely to engage in collective action on climate change or express a willingness to do so, even when controlling for several correlates of environmental behavior"
Evidence from Taliban-controlled Afghanistan shows that digital aid is a cost-effective, credible, and efficient way to reach vulnerable populations, in this case poor, tech-illiterate, female-headed households, in fragile states.