Although artificial intelligence is still a mature technology, it is already beginning to be used by election authorities in many countries. While most applications remain low-risk, a growing number of election management bodies (EMBs) are beginning to explore how to integrate more advanced forms of AI into election management throughout the election cycle and how to adapt their operations to the AI era. These new applications are not simply added onto existing processes. Rather, it is used to reimagine the core of election architecture and communications in ways that can significantly improve voter services and solve complex administrative challenges. At the same time, new risks may arise, especially when integrated into core election processes and public-facing applications such as voter roll management, vote tabulation and aggregation, and voter outreach. Concerns range from those related to the design of the AI itself (AI could misrepresent facts about the electoral process in ways that risk being biased, discriminatory, or disenfranchising voters) to concerns about path dependence and uneven institutional adoption, often due to limitations in AI literacy and capacity at the organizational level.
As EMBs expand their AI integration, risk is no longer hypothetical. In response, EMB is rethinking its strategy and updating its policies to ensure that AI can enhance election management without jeopardizing democratic principles. But election AI is still a largely uncharted field, so there are few precedents or best practices. This means that nearly every EMB has a unique approach to AI deployment and regulation. To capture these diverse institutional changes, International IDEA surveyed EMBs around the world and documented their different approaches to AI as an election technology.
International IDEA has teamed up with Microsoft, Arizona State University (ASU), and The Elections Group (TEG) to launch the AI + Election Clinic Skills Hub, part of the How Democracy Works Lab hosted by ASU. It is envisioned as a central knowledge repository of best practices and use cases to enable election officials to responsibly approach previously untapped AI-driven election technologies.
Skills Hub features case studies created by International IDEA and based on interviews with election authorities, covering unique AI use cases and governance structures in Estonia, Mexico, the Philippines, Pakistan, South Africa, Albania, Nigeria, the United Kingdom, and Australia. These studies, along with other materials from the Clinic, will serve as a capacity-building peer exchange resource for EMBs seeking to leverage AI innovation to reduce administrative burden, improve electoral processes, and redefine how EMBs run elections and how voters experience elections, while strengthening democratic principles. The case studies provide important lessons on how to strengthen AI readiness while avoiding pitfalls, helping EMBs learn not only how to leverage AI but also how they can develop internal procedures, staff programs, organizational structures, and AI policies to support the democratic integration of AI tools.
The Skills Hub is an extension of International IDEA’s extensive work on AI in election governance. This includes the Artificial Intelligence for Election Management report, which provides a knowledge base on the potential opportunities and challenges of using AI in election management, and the AI for Electors Workshop series, which was conducted in five countries and aimed to strengthen institutional capacity for the measured use of AI in elections. This website takes the next step and explores EMB’s real-world experience in AI integration.
While EMBs may share similarities in terms of current AI usage and governance structures, each case is different and highlights core ideas related to the practical realities of AI implementation. Albania and Australia provide two interesting examples that highlight the complex and unpredictable nature of AI deployment and highlight how a robust governance architecture and real-world testing are critical to responsible AI integration.
Albania
Albania’s Central Election Commission (CEC) follows an empirical approach when introducing new technologies into its operations. In line with this perspective, the CEC follows a testing protocol set for all new AI technologies, mandating rigorous prototyping and evaluation under close human oversight before solutions can be deployed at scale. During some 2025 mayoral elections, CEC tested a new AI-based image analysis tool for vote tabulation. The CEC’s experience with this tool, which was tested for the first time in a real election environment, provides an interesting example that highlights the importance of practical pilot testing.
The image analysis tool scanned ballots handled by poll workers to speed up vote counting and results transmission. As this was the first legally mandated pilot test, it was implemented in only 3% of polling stations and a manual counting process was also used in parallel. Although initially performing well with consistent accuracy, monitoring staff encountered two unexpected problems due to the physical conditions of polling places. First, the system tended to count votes twice if the ballots were held for too short a time. The problem became even more apparent late on election night, when tired poll workers held up ballots too quickly for the system to function properly. Second, the AI had difficulty recognizing even slight discrepancies in ballots, often resulting in valid ballots not being counted. The problem became especially noticeable after staff regularly handled the ballots, causing the marks on the ballots to fade, making the system less reliable over time.
This example illustrates the important lesson that many of the potential threats posed by new technologies cannot be fully understood before they are actually implemented. While pre-assessment and risk assessment are important steps in the development of AI tools, there are always unforeseen circumstances that are learned in the field that have a significant impact on the effectiveness and reliability of the system. For CEC, this experience reinforced the idea that prudent AI implementation requires proper prototyping and real-world testing so risks can be identified and addressed before they can cause significant harm to election integrity.
Australia
The Australian Electoral Commission (AEC)’s approach to AI integration is deliberately conservative, with core election management and voting conducted using strictly analogue methods. AEC is exploring AI tools that improve administrative efficiency, internal productivity, communications, and voter services while maintaining clear boundaries to protect the integrity of core processes. At the heart of these efforts lies careful consideration of not only the technical elements of application development, but also how institutional design and regulation need to evolve in parallel to establish support and safety guardrails for new AI tools.
To facilitate this prudent approach, the AEC is drafting internal AI guidelines and strategies in collaboration with the Australian Government’s expanding democracy stack governing AI in the public sector. This multi-layered system will need to be navigated by the AEC as it considers new AI use cases, spanning governance frameworks, assurance mechanisms, ethical principles, and technical standards. These governance instruments not only set rules and boundaries for AI use, but also provide resources to help public agencies build their capacity to leverage AI, such as training programs and government-owned sandbox environments to test AI tools without relying on third-party services. Such resources address the fact that many public authorities not only face a lack of policy frameworks governing the use of AI, but also lack the specialized technical standards, resources, tools, and expertise needed to realize AI projects.
In parallel with the government-wide AI policy, the AEC has established internal support structures to navigate this multi-layered governance framework and facilitate the development of rules-compliant AI tools. First, the Commission appointed two government-mandated oversight roles, including a Chief AI Officer to lead the AEC’s efforts in AI development and a Chief AI Officer to oversee the responsible use of AI across the AEC. Second, the AEC has established an AI Working Group, a dedicated forum to discuss, develop, and test ideas for new AI applications. This is open to all staff and serves as an important forum to ensure internal transparency in system development. Its broad remit includes providing guidance to staff on AI procurement, increasing AI awareness and literacy, and thereby ensuring a participatory implementation process.
Transparency is the central accountability mechanism throughout this system. Federal agencies are required to regularly update transparency statements explicitly detailing which AI tools they use, for what purposes, and under what conditions. By being subject to public scrutiny by design rather than choice, the AEC’s AI footprint will always serve the public and their democratic interests, reduce the risk of misaligned governance within the Commission, and add an additional layer of protection against problematic AI deployments. Australia’s AI governance stack, including transparency requirements, highlights how a layered policy infrastructure, rather than a single regulation, can strengthen accountability and implementation of AI by ensuring all stakeholders, including the public, are considered during the development of new applications.
Why this research remains important
The experience of the nine EMBs featured in the AI + Election Clinic Skills Hub is just the tip of the iceberg. In just a few years, many of the AI use cases featured in our case studies have gone from unimaginable to real. As AI technology becomes more sophisticated, AI adoption is likely to follow the same trajectory. However, these steps are not performed in the dark. EMB is acutely aware of the inherent risks associated with using AI in high-risk electoral situations, from systemic failures to discrimination and disenfranchisement. This recognition is clearly reflected in the universal shift from discrete and sporadic AI deployments to more developed AI strategies. An AI strategy not only establishes norms for how, when, and why AI will be used in election management, but also envisions the substantive ways in which AI can transform various aspects of EMB operations.
If risks are not adequately addressed through institutional adaptation, the integration of advanced AI capabilities into core election campaigns could not only cede direct control over certain processes, but also make the decision-making processes behind the outcomes produced highly opaque. This raises concerns about public trust and legitimacy, and undermines the integrity of elections and democracy as a whole. EMBs must therefore continue to explore new avenues for AI implementation, while also placing a clear focus on public transparency and creating clear pathways for public accountability and responsibility.
The AI + Election Clinic Skills Hub serves as a tool to facilitate this transparency-first approach. It is a place to share AI-driven innovation and, just as importantly, promotes institutional accountability by making EMBs’ AI experiences visible to their peers and the broader public. This is an important step for EMBs to gain democratic legitimacy for their decisions.
Far from prescribing a universalist approach to AI deployment, these case studies highlight how countries can devise their own AI stacks that simultaneously respect local norms and unique national contexts, while maintaining universal principles of ethical AI consistent with expert assessment, human rights, and democratic values. As EMBs continue to devise imaginative AI-based solutions to election problems, it is important that these ideas are collected and openly shared so that we can understand and learn from these potentially high-impact developments. This website will continue to collect these unique experiences and provide them as a platform.
If you’re interested in learning more about EMB’s experience with AI and how AI policies are being developed as guardrails, read our case study series in the AI + Election Clinic Skills Hub.
