Several groups have highlighted the risks of relying on AI-assisted systems in support of life-and-death decisions