It has now become a realistic prospect for smart care to be provided at home for those living with long-term conditions such as dementia. In the contemporary smart care scenario, homes are fitted with an array of sensors for remote monitoring providing data that feed into intelligent systems developed to highlight concerning patterns of behaviour or physiological measurements and to alert healthcare professionals to the need for action. This paper explores some ethical issues that may arise within such smart care systems, focusing on the extent to which ethical issues can be addressed at the system design stage. Artificial intelligence has been widely portrayed as an ethically risky technology, posing challenges for privacy and human autonomy and with the potential to introduce and exacerbate bias and inequality. While broad principles for ethical artificial intelligence have become established, the mechanisms for governing ethical artificial intelligence are still evolving. In healthcare settings the implementation of smart technologies falls within the existing frameworks for ethical review and governance. Feeding into this ethical review there are many practical steps that designers can take to build ethical considerations into the technology. After exploring the pre-emptive steps that can be taken in design and governance to provide for an ethical smart care system, the paper reviews the potential for further ethical challenges to arise within the everyday implementation of smart care systems in the context of dementia, despite the best efforts of all concerned to pre-empt them. The paper concludes with an exploration of the dilemmas that may thus face healthcare professionals involved in implementing this kind of smart care and with a call for further research to explore ethical dimensions of smart care both in terms of general principles and lived experience.