Handreichung 2/2025
Generative AI in Democracy Evaluation: Opportunities, Risks, and Ethical Considerations
Quito Tsui and Linda Raftree
Generative AI applications (so-called GenAI applications) have a significant impact on democratic processes and their evaluation. While AI has the potential to improve access to and evaluation of these processes, it also poses risks in terms of destabilizing and eroding them. This Handreichung examines the role of AI in democratic systems and its application in the areas of monitoring, evaluation, research and learning.
The authors analyze the key issues, challenges and ethical concerns. They discuss the potential uses of AI, argue for its responsible use and derive recommendations for various stakeholders – including policymakers, practitioners and evaluation societies.
About the Authors
Quito Tsui is a researcher and writer with expertise on technology in the context of the humanitarian and development sectors. She has a keen interest in the development of a technological environment rooted in justice with practical experience of supporting organisations to this end. As a Core Collaborator at the MERL Tech Initiative, and Humanitarian AI + MERL Working Group lead she has conducted original research on deduplication, and developed guidance on safe and responsible use of AI in social impact efforts, humanitarian applications and research on violence against women. She previously led research on digital identity, biometrics, interoperability and environmental justice at The Engine Room. She has also conducted academic research at the University of Cambridge, Stanford University and the London School of Economics. Quito’s writing has been featured in the Forced Migration Review (Oxford University), Bot Populi and The Nation and her work has been cited by outlets such as Access Now and Geographical.
Linda Raftree focuses on the ethical use of technology and data in international development, human rights, and humanitarian contexts. She is the founder of the MERL Tech Initiative (MTI) which has been exploring the responsible use of digital tools and data since 2014. Through MTI, Linda has played a strong role in bringing together critical thinkers from the monitoring, evaluation, research, and learning (MERL), data science, and digital technology sectors to collaborate, develop capacities and strengthen good practice. Linda convenes the MERL Tech Natural Language Processing Community of Practice (NLP-CoP) and the New York City Technology Salon. She has advised a range of foundations, bilateral agencies, and non-governmental organizations on responsible data policy and practice, applied research, digital safeguarding, inclusive digital approaches, and safe programming. Linda is a Certified Information Privacy Professional (CIPP & CIPM).
