One of today’s most fascinating challenges is understanding how to harness artificial intelligence (AI) to reduce social isolation, which is a public health problem whose damages have been compared in scale with cigarette smoking. Social isolation is well-known to accelerate dementia, as well as death. How can we direct the deep thinking and learning of AI to reduce social isolation?
The depth of thinking and learning used to create AI is an inspiration for calling AI scientists to also think deeply and reflect upon the essence of social isolation, which is defined as the objective paucity of meaningful and positive social contacts with other human beings, measured in terms of network size and community involvement (see also Machielse, A., 2006; and Victor, C., et al., 2000).
Deep thinking in AI refers to the ability of systems to process, analyze, and derive insights from vast amounts of data in a manner that simulates human cognitive processes. Deep thinking is primarily powered by deep learning, which is a specialized branch of machine learning within the broader field of AI. Deep learning uses multilayered neural networks—hence the term “deep”—to analyze data and recognize complex patterns. This allows computers to perform tasks such as image recognition and natural language processing with remarkable accuracy.
But AI will only acquire the ability to decrease social isolation once AI researchers deeply and fully understand the phenomenon of social isolation. A major barrier standing in the way of AI’s ability to tangibly decrease social isolation is the limited understanding AI researchers have about social isolation. Evidence of AI researchers’ limited understanding about social isolation is an increasing body of publications about AI applications aimed at addressing social isolation in a superficial way.
‘AI will only acquire the ability to decrease social isolation once AI researchers deeply and fully understand the phenomenon of social isolation.’
One hallmark of this superficial understanding of social isolation is that AI technologies are presented as a healthy alternative to interacting with other human beings, with the underlining assumption that interacting with things manifesting as screens, computers and robots is as beneficial as interacting with human beings. Perhaps this limited understanding stems from AI researchers’ limited collaborations with social scientists who are experts on social isolation, as well as limited time spent with older adults who are socially isolated, hence difficult to identify, let alone spend time with.
To offer a few examples of AI technologies “combating” social isolation, AI researchers tested a tiny robot chatting with older adults with dementia who live alone in Osaka. Researchers from New York University applauded messages from ChatGPT such as: “if you ever need someone to talk to, I am here to listen and support you,” as a tool to relieve older adults’ isolation.
Similarly, researchers at Harvard presented messages such as “I need you,” coming from a chatbot as an efficient strategy to address loneliness. Students from Johns Hopkins University introduced an online game in a retirement community near their school to reduce isolation in its residents. Missing from these well-meaning initiatives are deep reflections on the possibility that offering a robot, a screen, a tablet or some other lifeless object as a way to interact more with others could be an innovative and covert way to perpetrate, and perhaps even exacerbate, social isolation, which is the limited access to meaningful and positive interactions with other human beings.
To truly address social isolation, we invite AI researchers and their funders to think bigger, bolder and much deeper. One example of expansive thinking comes from Mark Zuckerberg, who recently shared his vision that superintelligence, a form of advanced AI, will empower “people to be more creative, develop culture and communities, connect with each other, and lead more fulfilling lives.”
Along those lines, questions that might aid how AI researchers think about eradicating social isolation among older adults could include: how can AI foster meaningful and positive interactions among users and other human beings? How can AI foster social integration, community involvement, and social justice? These questions also act as an invitation to move AI researchers’ perspective (as well as funders of their research) beyond the personal sphere of potential users toward systems used by healthcare professionals and officers at the city, state and federal levels to support older adults who are socially isolated.
‘A common thread in all these uses of AI is that AI is the bridge from isolated older adults to other human beings.’
Examples of deeper and smarter applications of AI to address social isolation can include strategies to identify and map socially isolated older adults, with the goal of integrating them socially (if this is their wish), plus protecting them in event of natural disasters. Other applications could include ways by which older adults who are isolated due to living in remote areas can easily interact via video with other human beings for company, for finding important information, and for being visited by a healthcare provider.
Yet another possibility is to hone AI applications to self-diagnose chronic conditions such as cognitive impairment and brainstorm ways to cope with these conditions, especially when living alone, which is a likely sign of social isolation. AI applications also can jog the memory and thinking abilities of people with cognitive impairment to help them interact easily with other human beings. And AI can help rehearse difficult future conversations with family members and healthcare providers. A common thread in all of these AI uses is that AI is the bridge from isolated older adults to other human beings, such as physicians, family members, friends, and society at large.
AI can foster social integration across generations and communities because AI can make us more informed, connected, tech-savvy and conscious. An important component of this consciousness is to celebrate, design, fund, and enjoy AI technologies connecting human beings, with attention to those who are most isolated.
We call for research testing AI applications’ effectiveness in increasing meaningful social connections among human beings. A reflective, interdisciplinary, and culturally humble approach is critical to harness the immense benefits AI can offer: the sky is the limit.
Elena Portacolone, PhD, MBA, MPH, is a professor at the Institute for Health & Aging, University of California, San Francisco.
Photo credit: Shutterstock/WOSUNAN
