Time-shifting or the ‘retrograde plunge’ is a common phenomenon in dementia, particularly Alzheimer’s disease, where a person experiences events, roles, and relationships from their past as existing in the present day. This may involve a desire to speak to a deceased spouse or wanting to leave a care facility to go to work. When these requests are not accommodated, the person with dementia may become increasingly agitated or distressed. In these circumstances, and to act in the perceived best interests of a person with dementia, it is common for care staff to use distraction or a therapeutic lie to manage distress.
The application of AI is increasingly being considered across various domains in dementia care. Using the example of a consent-driven and co-constructed chatbot, this presentation will sketch the ethical terrain of how AI may be used to engage a person when in a distressing retrograde plunged state. As with therapeutic lies this approach aims to ease distress for a person with dementia. In addition, it may also potentially facilitate experiential coherence with a retrograde plunged state, maintain elements of personal narratives and identity, and offer professional carers insight into a person’s biographical history to support personal understanding and enhance future therapeutic engagement. Many of the ethical considerations that guide the use of therapeutic lies will be applicable. However, the potential for AI to make the retrograde plunge ‘immersive’, such as impersonation of loved ones, will require ethical consideration beyond that of therapeutic lies.