In this reflection from the Data for Public Goals event at Politecnico di Milano, we call for a radical reframing of how we understand and govern data for AI. Moving beyond technical metrics and ethical tweaks, the piece introduces a planetary justice lens that traces AI's impacts across its full supply chain–from resource extraction to model deployment. Highlighting the ecological, labor, and epistemic costs embedded in AI systems, the post challenges the narrative of data neutrality and questions the assumption that automation is inherently beneficial. What if, instead of optimizing AI, we asked whether it is needed at all–and who gets to decide?

The AI + Planetary Justice Alliance joined researchers, organizers, and practitioners in Bengaluru for the Mapping Data Work workshop — a gathering that foregrounded the invisible labor behind AI systems. In our blog, we report key takeaways, reflect on what it means to build “ethical datasets,” question the assumptions behind datafication, and share insights from our own research on epistemic justice in agricultural AI in India. What if the push to represent is just another form of surveillance? And why are we building AI in the first place? Read the full piece to explore the planetary justice implications of some of AI’s quietest supply chains: data work.