The inadequacy of the dementia care workforce in meeting the needs of a growing population of people living with dementia (PLwD) is a global public health challenge. There is a burgeoning interest in exploring the role of artificial intelligence‐enabled technologies, such as chatbots, neural network models, and other programs designed to intelligently simulate human conversations, in the dementia care landscape. The aim of this scoping review was to examine the themes and gaps associated with the use of these technologies in the dementia care space.

Three databases (PubMed, PsychINFO, and CINAHL) were searched for English‐language, peer‐reviewed articles using keywords relating to chatbots, virtual assistants, generative pre‐trained transformer technology (GPT‐3), and dementia. The search returned 26 articles after removal of duplicates with 13 selected for inclusion into the review.

Two user groups were identified, PLwD and their family care partners. Chatbots, embodied conversational agents, and intelligent virtual agents were the identified technologies. The core uses of the technology centered around 1) diagnosis and monitoring of cognitive symptoms; 2) social connection and cognitive simulation; 3) memory aids and prompts for task completion, and 4) education, training, and social support for care partners. Five key themes emerged, including 1) the need for expanded studies that go beyond superficial and simple dialogues to more complex and realistic interactions; 2) the importance of a person‐centered approach that uses knowledge about the person to guide the technology; 3) small sample sizes and weak quality across studies; 4) the need for standardized criteria to evaluate the quality of human/technology interaction; and 5) the importance of co‐creation of technology with insights from PLwD and their care partners.

With the growing capability to generate human‐like text for powering chatbots, conversational artificial intelligence has the potential to play an important role in the dementia care space, however the literature remains in its nascent stages. There is a need for well‐designed studies, utilizing multidisciplinary and participatory approaches, exploring the impacts of incorporating these technologies into the daily lives of PLwD and their care partners.