How can we build ethics into artificial intelligence (AI) research? That is the question posed by researchers connected with grantee Partnership on AI. A paper published in Nature Machine Intelligence offers recommendations on ethics review and including diverse voices in and beyond the AI community. Current practices for ethics review on journal article submissions vary from publication to publication, with most journals lacking an ethics review process. The authors call for a cultural shift in practice that moves towards greater community coordination and shared norms that seek to anticipate and mitigate the potential harms of AI research.