Singapore AI Governance Framework MAF-3.7(1): Allocate clear, trained and resourced roles for ethical AI deployment

MAF paragraphs 3.4 to 3.7(1): adapt existing governance structures (for example managing AI risk within enterprise risk management and ethics through an ethics review board), centralised or decentralised as fits, with top management and board sponsorship; allocate responsibility for each stage of AI deployment, possibly through a cross-organisation coordinating body; make sure people with AI governance roles are trained and resourced; key roles cover assessing risks and impacts on individuals, deciding the level of human involvement, managing model training and selection, maintaining, monitoring, documenting and reviewing deployed models, reviewing communication and feedback channels, and training staff who work with or answer questions about AI systems.