With a lot to play for, including legal precedents, how can dev teams approach genAI risks with a view to strong mitigation strategies?