Abstract
<jats:p>Background: Dementia affects more than 57 million people worldwide and remains one of the most disabling and least treatable syndromes associated with aging. Given the lack of a pharmacological cure, there is significant interest in utilizing artificial intelligence to improve both the accessibility and quality of care. At present, AI-powered tools are applied in three main areas: early diagnosis through neuroimaging analysis, non-pharmacological management of behavioral symptoms with social robots, and continuous safety monitoring using smart home and wearable technologies. Objective: This review synthesizes peer-reviewed evidence from PubMed published between 2015 and 2025, evaluating the practical impacts of AI tools in dementia care and examining the moral challenges associated with their use in populations experiencing progressive cognitive decline. Key Findings: Deep learning techniques applied to brain MRI have demonstrated diagnostic validity for Alzheimer's disease that equals or exceeds the performance of specialist clinicians. However, questions remain regarding their effectiveness across diverse populations. Robotic companions, such as the PARO seal robot, have been shown to reduce agitation and depression in individuals with dementia and to significantly decrease the use of psychotropic medications. Observation technologies improve safety and alleviate caregiver anxiety, but they also introduce surveillance practices that may compromise the individual dignity and autonomy of those they are intended to protect. Conclusions: AI provides significant clinical benefits in dementia care; however, these advantages are inseparable from the ethical conditions governing its use.</jats:p>