Risk-based approaches to governance bear an ambiguous stance regarding the Research and Development stages of AI, for they the possibility of explicit risks before they are posed by a given finalised product. In this context, Institutional Review Boards (IRBs) stand as unique governance mechanisms, capable of addressing the step from general research to concrete product development. However, IRBs face several challenges in governing AI-based medical products, including: (a) achieving consistency, (b) being exhaustive, (c) ensuring process transparency, and (d) reducing the existing capacity and knowledge asymmetry between different stakeholders. This article explores four governance levers that can be used to effect change, four governance entry-points throughout a product's lifecycle, and five different behaviours that IRBs should try to advance to ensure the effective governance of the R&D stages of AI-based medical projects. In doing so, IRBs can seize the unique opportunity they present to bring principles into practice, increase research quality, reduce governance costs, and bridge the knowledge gap between stakeholders.

翻译：基于风险的治理方法在人工智能的研发阶段持有一种模糊立场，因为它们难以在最终产品构成明确风险之前识别这些风险。在此背景下，机构审查委员会（IRBs）作为一种独特的治理机制脱颖而出，能够应对从基础研究到具体产品开发的跨越。然而，IRBs在治理基于人工智能的医疗产品时面临若干挑战，包括：（a）实现一致性，（b）确保全面性，（c）保证流程透明度，以及（d）减少不同利益相关者之间现有的能力与知识不对称。本文探讨了可用于推动变革的四种治理杠杆、贯穿产品生命周期的四个治理切入点，以及IRBs应努力促进的五种不同行为，以确保基于人工智能的医疗项目研发阶段的有效治理。通过这种方式，IRBs能够把握其独特的机遇，将原则付诸实践，提高研究质量，降低治理成本，并弥合利益相关者之间的知识鸿沟。