In an effort to share ideas beyond the limited time we have during the annual conference, the Community Building Committee asked Graduate Students attending TPRC54 to take notes of paper sessions to be made available to the community. The following is a summary of the paper session entitled “AI and the Public” that was presented on Saturday, September 26, 2026 at 11:00am. The session was moderated by Gwen Shaffer, California State University Long Beach. The following is an edited summary based on the notes provided by Ashley Timean, Carolina University. Full papers can be accessed through ConfTool.
Paper 1: AI Localism - Counter Imaginaries from City AI Governance
Bhavana Bheem, Massachusetts Institute of Technology and Samuel Gerstein, University of Illinois Urbana-Champaign
This presentation examined how city and county governments across the United States are developing policies to govern artificial intelligence (AI), particularly in the absence of comprehensive federal legislation. Local governments are increasingly becoming frontline AI regulators, a trend the researchers describe as AI localism. The study analyzed 77 AI policy documents from 55 local governments across 25 states, Washington, D.C., and Guam.
The findings showed that responsible and ethical AI use is a priority for local governments of all sizes, with policies addressing transparency, accountability, consent, human rights, bias, and restrictions on AI use in high-stakes decisions. However, priorities varied by community. Larger cities focused more on automated decision-making and technological development, while smaller cities placed greater emphasis on disclosure and alignment with community needs. The researchers identified three common approaches to local AI governance: democratic oversight, which gives residents the ability to understand and challenge government uses of AI; rights-oriented governance, which connects AI policies to existing privacy and civil-rights protections; and place-based alignment, which ensures AI systems reflect local problems and community values.
Overall, the presentation demonstrated that AI governance is not shaped solely by technology companies and national governments. Cities, counties, and their residents are already influencing how AI is used and regulated, creating approaches that reflect their communities' priorities, experiences, and lived realities.
Paper 2: Generative AI as the Third Crisis of the Press: Inequality, Trust, and the Future of Publics
Juan Ortiz Freuler, University of Southern California and Bumju Jung, University of Southern California
This presentation examined the three major systemic crises the press has experienced over the past 30 years as technology has changed how news is gathered, verified, packaged, and distributed. The presenter emphasized the press’s traditional role in informing the public, which is essential to legitimate democratic decision-making. Each crisis has disrupted a different stage of the press’s production model. Beginning around 2005, the shift from print to the web weakened news organizations’ control over how their content was packaged and contributed to the loss of advertising revenue to online platforms. Around 2017, social media platforms gained greater control over news distribution and access to readers. Now, beginning around 2022, generative AI has created a third crisis by allowing machines to reproduce content at unprecedented speed and further challenging the press’s control over content production.
The presenter identified three challenges generative AI poses to journalism and the public. First, inequality of access is reinforced by high costs, economies of scale, and network effects that have concentrated AI services among a small number of providers. Second, synthetic interactions can disrupt public discourse. Bots can impersonate people, create the appearance of public consensus, and discourage genuine democratic deliberation. Third, generative AI can erode trust by fabricating articles, URLs, and sources, while undisclosed training datasets make independent auditing difficult. Together, these challenges threaten journalism’s ability to serve as a trustworthy information source and fulfill its democratic watchdog role.
In response, the presenter proposed three emerging principles for AI governance: fair access, protecting human agency, and transparency. Fair access means expanding access to data, computing power, and generative AI models. Protecting human agency requires meaningful human oversight and ensuring that people retain final authority over decisions. Transparency means allowing audiences to trace claims to their sources and recognize when automation has been used. The presentation concluded that generative AI represents the press’s third systemic crisis of the twenty-first century. While the challenges are unprecedented, lessons from earlier policy discussions remain relevant, and emerging policy approaches offer some hope for addressing these threats.
Paper 3: Civil Society and Agility in AI Governance
Ganchimeg Namsrai, Department of Media and Information, Michigan State University; Johannes M. Bauer, Quello Center, Michigan State University; Ruth Shillair, Department of Media and Information, Michigan State University
This presentation examined how civil society can contribute to more agile and ethical AI governance amid rapid technological development, technical complexity, uncertain future effects, and competing stakeholder interests. The presenter argued that effective AI governance is unlikely to emerge from institutions working in isolation. The research brought together two areas often considered separately: the need for adaptable AI governance and the role of civil society in shaping it. The study focused on two questions: how civil society participation is organized and how that participation contributes to more agile AI governance.
The researchers examined civil society involvement at the international, federal, and state levels, as well as the activities of civil society organizations in the United States. Internationally, United Nations documents recognize the importance of civil society, and UNESCO has established a dedicated network for civil society and academia. Other international organizations include civil society through consultations, summits, and working groups, although the form and influence of that participation are not always clear. At the U.S. federal level, participation is still developing, with no consistent formal mechanism for including civil society in AI governance. At the state level, researchers identified eight enacted laws that formally included civil society representatives in task forces, commissions, advisory groups, and working groups. Civil society representatives held approximately 6% to 35% of the seats on these bodies, with labor, civil-liberties, privacy, and disability-rights organizations most frequently represented. However, most of these bodies were advisory, focusing on meetings, public hearings, assessments, and reports or recommendations.
The presentation also highlighted that limited formal representation does not mean civil society organizations are inactive. These organizations research and compare AI policies, identify concerns such as bias, raise public awareness, train policymakers and researchers, and bring stakeholders together to develop practical governance frameworks. They contribute important contextual knowledge, evidence, lived experience, diverse perspectives, and public-interest values. However, their ability to participate varies considerably depending on their size, resources, funding, reach, and independence.
The presenter concluded that representation alone does not guarantee meaningful participation or influence over policy. For civil society organizations to contribute effectively, they need sufficient capacity, independence, access to relevant information and data, and institutional arrangements that support their involvement. Because it is still too early to determine their overall impact on AI governance, further research is needed to understand which forms of participation are most effective and what institutional conditions allow civil society to help shape agile and ethical AI governance.
Session Reflection
“I enjoyed how the three presentations examined AI governance from different but connected perspectives. Presentation 1 showed how local governments are developing AI policies that reflect community values, rights, and immediate concerns. Presentation 2 explained how generative AI represents a third systemic crisis for the press by disrupting content production, public deliberation, and trust in information. Presentation 3 demonstrated how civil society can contribute knowledge, diverse perspectives, and public-interest values to more agile and ethical AI governance. I learned that AI governance is not only the responsibility of national governments and technology companies. Local governments and civil society organizations can influence how AI is regulated, while the press remains essential for helping the public understand and respond to these changes. I also learned that transparency, democratic oversight, human agency, fair access, civil rights, and community participation were common themes across the three presentations.
My main question is how communities, journalists, and civil society actors can be given sufficient authority and resources to hold powerful AI developers accountable.”