EAGER: AI-DCL: Addressing sociotechnical challenges of conversational agents and interventions in the context of elderly care

研究人员将开发和评估一种AI对话代理，以帮助认知能力下降的老年人。该代理将使用对话模式开发，旨在解决用户的理解和接受问题;它们将用于协助日历和事件相关任务。在原型药剂的实验室开发之后，将对原型的家庭使用进行评估。开发人工智能技术以支持那些在当地养老的人是非常有必要的，这个项目有创新的计划。这些影响将通过与相关利益相关者群体的合作来实现，包括对话代理的设计者和开发者，帮助老年人的医疗保健专业人员以及老年人本身。更广泛地说，如果该项目成功地帮助老年人适应使用会话代理来协助完成与日历相关的任务，那么这一成功可能会帮助他们更适应使用人工智能代理来协助与身心健康相关的其他活动。简而言之，该项目的结果可能有助于解决随着时间的推移不断上升的医疗保健成本。该奖项支持一项研究项目，该项目使用人类-集中的设计方法来整合生理和会话数据，以评估用户的身体，情感和认知状态。然后，这种评估将被考虑到基于对话的语音激活系统如何响应用户。研究团队将为会话助理开发一个基于话轮的对话框架。该框架将围绕处理日历和活动，如社交和休闲活动，或老年人的健康相关预约。然后，它将用于开发一个支持日常活动调度的原型AI系统。该团队还将开发个性化的用户模型，在考虑如何最好地与老年用户互动时，将认知和身体能力考虑在内。他们还将研究老年人或其他人更喜欢与会话代理互动的长期变化，因为他们对它越来越熟悉。该项目的结果将有助于改善因疾病或高龄而能力下降的人的健康，福祉和独立性，该奖项反映了NSF的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

The investigators will develop and assess an AI Conversational Agent that will assist elderly persons with cognitive decline. The agent will be developed using a dialogic model which is intended to address user issues of comprehension and acceptance; they will function to assist with calendar and event-related tasks. Laboratory development of prototype agents will be followed by assessment of in-home use of the prototypes. There is a very substantial need to develop AI technologies to support those ageing in place, and this project has innovative plans for doing so. These impacts will be achieved by engaging with pertinent stakeholder groups including designers and developers of conversational agents, health care professionals who assist the elderly, and the elderly themselves. More broadly, if the project succeeds in helping the elderly become comfortable with using conversational agents to assist with calendar related tasks, that success may help them to become more comfortable with using AI agents to assist with other activities associated with physical and mental health; in short, the results of this project may serve to address rising health care costs over time.This award supports a research project that uses a human-centered design approach to integrate physiological and conversational data to assess the physical, emotional, and cognitive status of the user. That assessment will then be factored into how a conversation-based voice-activated system responds to the user. The research team will develop a turn-based dialog framework for a conversational assistant. The framework will center around handling calendars and activities, such as social and leisure activities, or health-related appointments for elderly people. It will then be used to develop a prototype AI system that supports daily activity scheduling. The team will also develop personalized user models that factor in cognitive and physical ability when considering how best to interact with elderly users. They will also examine longer-term changes in how older adults or others prefer to interact with a conversational agent, as they gain more familiarity with it. The results of this project will contribute to improving the health, wellbeing, and independence of people who have reduced capacity due to illness or advanced age, or simply who live alone.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.