AI Systems as Digital Public Goods

This research report by the United Nations University and the Asian Development Bank explores the evolving framework of Artificial Intelligence as Digital Public Goods (AIDPGs). It evaluates how traditional standards for open-source software must adapt to address the unique complexities of AI systems, such as training data transparency, model weights, and long-term safety. The report highlights that openness is a spectrum rather than a binary state, requiring a balance between accessibility and ethical safeguards to prevent harm. Experts advocate for modular governance that monitors AI throughout its entire lifecycle rather than only at the point of release. Ultimately, the sources provide a strategic roadmap organized around the SAFE framework to ensure AI development fosters global equity and aligns with Sustainable Development Goals.