The European Commission's Green Deal is an opportunity to rethink harmful practices of research and innovation policy.

The European Union's Green Deal and associated policies, aspiring to long-term environmental sustainability, now require economic activities to 'do no significant harm' to EU environmental objectives. The way the European Commission is enacting the do no significant harm principle relies on quantitative tools that try to identify harm and adjudicate its significance. A reliance on established technical approaches to assessing such questions ignores the high levels of imprecision, ambiguity, and uncertainty-levels often in flux-characterizing the social contexts in which harms emerge. Indeed, harm, and its significance, are relational, not absolute. A better approach would thus be to acknowledge the relational nature of harm and develop broad capabilities to engage and 'stay with' the harm. We use the case of European research and innovation activities to expose the relational nature of harm, and explore an alternative and potentially more productive approach that departs from attempts to unilaterally or uniformly claim to know or adjudicate what is or is not significantly harmful. In closing, we outline three ways research and innovation policy-makers might experiment with reconfiguring scientific and technological systems and practices to better address the significant harms borne by people, other-than-human beings, and ecosystems.