In response to the limited knowledge and understanding of what might happen in a geoengineered world, some call for an expansion and acceleration of research (4), including stepping beyond modelling studies to the design and conduct of field experiments or even proof of concept engineering trials. By definition, such experiments would need to be carried out at scales sufficient to generate measurable effects that could be distinguished from background variability, and the area (or volume) of impact could neither be precisely defined nor contained. This raises the questions of whether and, if so, how such proposed research could be properly assessed, regulated, controlled and monitored. How could research at scale be distinguished from actual deployment of a geoengineering technique? And who would bear responsibility for the review and authorisation of such research and would ultimately be liable for any damages or other impacts caused?