Adoption of Generative AI and Large Language Models in Danish Public Administration

This pilot study examines how employees and managers in Danish public administration experience the use of generative artificial intelligence and large language models in their work. Through workshops and interviews in two municipalities, the study identifies opportunities and challenges in adoption. Participants highlight benefits such as improved efficiency, enhanced text quality, and support for administrative tasks. However, adoption is shaped by legal uncertainty, data security concerns, unclear managerial guidance, and subtle social expectations to use AI efficiently. Use is often self-initiated and driven from the bottom up, creating a disconnect between government and local management aspirations and actual practices. In some cases, AI use remains informal or hidden due to concerns about compliance or perceived norms. These findings suggest that while AI offers real potential, successful implementation depends on more than technical access—it requires organisational readiness, legal clarity, and an awareness of the informal dynamics that shape everyday use.