Paper submissions are open at EasyChair
Co-located at WebSci 2025, May 20-24, 2025 in New Brunswick, New Jersey, USA



Why Elections? Why Now?
Elections represent a cornerstone of the democratic process, yet the intersection of the modern information ecosystem and electoral processes have revealed new and amplified pre-existing vulnerabilities in these processes. Much of the security focus around these electoral processes comes in the immediate lead-up to an election, yet many of these vulnerabilities—e.g., anxieties around voter fraud, perceptions of election integrity, partisan animosity/polarization—are present, grow, and could be better addressed between election cycles. These factors include direct vulnerabilities—-such as cyber security infrastructure and cyber-physical systems—and socio-technical issues that erode public trust in electoral processes or drive anti-social outcomes like increased polarization.
In the face of global democratic backsliding and rising populism, methods to improve security—and perceptions thereof—in these democratic processes while also supporting generally underfunded (and of- ten under-protected) state and local election officials represents a major challenge. In 2025, multiple major democracies will be holding elections, including Canada, Chile, the Philippines, Germany, and Ireland, with the United States launching multiple state gubernatorial elections and preparing for the midterms in 2026. Studies in this space are of timely, pressing societal concern, and with recent changes in policy and ownership of online platforms, new approaches to collection, analysis, and intervention are needed.
This workshop therefore facilitates the following objectives:
- Bring together interdisciplinary scholars to address issues of free, fair, open, and trustworthy elections and campaigns in the era of agentic AI;
- Learn from and share perspectives about how stakeholders have addressed election-related issues across both state- and national borders;
- Move beyond descriptions of misinformation, disinformation, and platform policies on election security by proposing solutions and frameworks for improving security, resilience, and civility; and
- Identify resource gaps and offer policy recommendations and practical tools to support state and local election officials in safeguarding elections in a challenging global environment.
Workshop Topics
The workshop focuses on two elections-related areas: AI-mediated political communication/behavior that have the potential to influence election outcomes, and securing the infrastructure on which elections and campaigns occur. Topics across these areas include:
- Interventions that improve democratic discourse (e.g., reducing affective polarization, stimulating engagement in civil political discussion);
- Assessing the impact of social media on democratic institutions and informed publics;
- Combating or increasing resilience to malevolent influence efforts (e.g., disinformation campaigns);
- Identifying and protecting vulnerable audiences (minoritized groups, groups vulnerable to radicalization, groups with limited internet skills) in the age of AI;
- Generative AI and its impact on trust of information;
- The role of AI in online political campaigning; and
- Frameworks for assessing interventions for improving election security.
Workshop Format
The workshop will contain two keynote talks (30 mins each), one oriented toward the role of AI in political communication and how new advances in agentic AI can support democratic engagement, institutions, and processes; and a second focused more on the security aspects of the election process. These keynotes will establish the interdisciplinary nature of this workshop, where we bring together social scientists from political science and communications, computational scientists from machine learning and information studies, and industry experts who develop tools for image content moderation.
In addition to these keynotes, we will have a series of contributed talks that highlight the best paper submissions and will hold a panel discussion on open challenges in characterizing the variety of images used in online spaces and limitations of these approaches when applied across multiple online spaces. We intend to host a data challenge and related poster session for submissions to describe their work in characterizing images shared in online spaces and identifying potentially anti-social imagery online.
Workshop Agenda
- Morning Session 1: Keynote # 1
- Morning Session 2: Paper Presentations
- Afternoon Session 1: Keynote # 2
- Afternoon Session 2: Data Challenge Results
- Afternoon Session 3: Panel and Closing
Publications
We will explore the possibility of publishing the best papers in a journal special issue. Potential journals include:
- ACM Transactions on Social Computing
- Social Media + Society
- Journal of Online Trust and Safety
- Computational Communications Research
- Journal of Online Social Networks and Media