This article discusses the way to develop personnel toward the AI era. With the spread of AI technology, we have to switch the way of personnel development, from the way that is based on instruction documents, to the way that is based on interaction among workers. To discuss this switch, we introduces the case of Japanese nursing homes that implements “learning treatment” to treat patients of dementia. This case considers how boundary crossing enhances collaborative learning in communities of practice (CoPs). CoPs were presumably connected horizontally and vertically, which could enhance boundary crossing by connecting several CoPs. Nursing homes that successfully established the “learning treatment”, a non-pharmacological treatment for dementia, are tapped to build CoPs that can learn and discuss knowledge and skills related to the treatment. The staff discovers new values of the CoPs, as well as acquiring new care skills. To improve treatment knowledge, staff conducts study visits to other nursing homes and participated in regional study meetings and learning treatment symposia representing workplace, company, regional, and prefectural boundary crossings. We discuss several aspects for building multi-layered CoP structures.
