AI’s role in healthcare has been hyped both via utopian and dystopian narratives, with the latter often used to argue that the technology is the only saviour to a labour and cost crisis. The problem with hype is that it undermines the nuance and debate needed if we are to use this still-largely-speculative technology for societal good. Where such technology is dominated by big firms based far away, how can municipal care providers ensure sufficient control to maintain quality when using AI? What procedures and processes are necessary to ensure that the use of AI in, for example, care assessment decisions doesn’t erode basic values and democratic rights within the welfare model? In considering these questions, the talk will outline the key political and ethical issues faced when adopting AI in healthcare with a particular focus on the welfare model.