Imagine the story of Prometheus, the Titan who gave fire to humanity. Fire transformed civilization - it lit our homes, fueled innovation, and enabled progress. Yet, fire also burned, destroyed, and caused harm when misused. Prometheus' gift was both a blessing and a curse with its impact depended not on the fire itself, but on how humans wielded it.

Artificial Intelligence is the fire of our era: powerful, transformative, and full of promise. But in the public sector, its impact hinges not on the technology itself, but on the manner of its adoption. Responsible AI is therefore not only about whether systems are “ethical by design,” but about governance, i.e., whether public administrations adopt AI in ways that uphold legality, accountability, fairness, sustainability, and inclusion. Without such governance, AI adoption risks becoming our “monster”: amplifying bias, eroding trust, and undermining democratic legitimacy.

This keynote explores that duality. Drawing on my ongoing research, we will unpack the most pressing data-related challenges to responsible adoption, highlighting the institutional and governance “symptoms” that signal high-risk deployments. Particular attention will be given to public data ecosystems as both enablers and bottlenecks of responsible adoption, with challenges mapped across data lifecycle phases and data roles. The keynote will conclude by outlining a forward-looking agenda and inviting reflection on how governments and researchers can strengthen public data ecosystems to ensure AI adoption fulfills its promise.