
Technology
Devin Liddell & Sheng-Hung Lee
Devin Liddell
Principal Futurist
Sheng-Hung Lee
MIT Ph.D. Researcher
Most of us want to remain in our existing homes as we grow older. The practice of “aging in place” aligns with preferences for familiar places and routines and preserves our sense of independence. These preferences, though, raise questions about what support seniors want and need in their current homes. Japan has advanced the use of robotics specifically for this purpose, with mixed results. Despite these early results, the continued development of robotics and artificial intelligence (AI) to assist those aging in place seems obvious. What’s less obvious is how seniors foresee AI and robots living alongside them and what specifically they envision these things doing.
To begin answering these questions, the Massachusetts Institute of Technology (MIT) and Teague developed a preliminary conceptual framework for understanding the future of AI and robotics in older people’s homes. We then used that framework to explore the aspirations—and concerns—of this population through semi-structured interviews with the MIT AgeLab’s research cohort of seniors. The result is a landscape view of potential future AI forms along with the key considerations we’ll need to account for when designing them.The x-axis of the conceptual framework distinguishes between digital and physical presences of future AIs in the home, while the y-axis differentiates between cognitive capabilities and tactile proficiencies. In other words, the left side of the framework features digital-only AIs, the right side of the framework features AI-powered robots, and the bottom of the framework is extra muscle for help with more tactile things while the top of the framework is extra brain power for help with more cerebral things.
Altogether, this conceptual framework outlines four major archetypes: an Advisor AI, a Butler Robot AI, a Conductor AI, and a Valet Robot AI. Here’s what those archetypes would do for future seniors at home, along with essential feedback from present-day seniors.
A digital presence that suggests solutions to problems, surfaces opportunities, and helps its person remember to do things. Examples: the AI helps verify the veracity of an unfamiliar communications like scam phone calls; identifies activities of interest and assists in planning how to participate; offers timely reminders to take medications, and prompts calls to friends and family members on their birthdays.
A physical presence that attends to its person by assisting with dynamic needs, such as deliveries, health and home monitoring. Examples: the AI robot lifts a delivery from the porch to the foyer; assists in turning off the water at the source of a leak in the kitchen; renders assistance—and summons help, if needed—in the event of a fall.
A digital presence that operates connected systems of modules such as wheeled porters and object lifters. Examples: the AI responds to voice commands to transport meals from the kitchen to the living room with a wheeled porter; elevates an adjustable-height table adjacent to the dryer to ease folding clothes; summons an autonomous vacuum to address a spill.
A physical presence that attends to its person by helping with everyday tasks, such as cleaning, dressing, and grooming. Examples: the AI robot replaces a light bulb in high-ceiling recessed lighting; helps a person put on their socks and pants; cleans everyday surfaces such as kitchen and bath countertops, and dusts bookshelves and framed prints.
As a landscape view of potential future AI forms in the homes of seniors, the conceptual framework and its four archetypes reveal some inherent blurriness across its vectors. For example, the aforementioned hazy boundary between a Valet Robot AI focused on helping with everyday tasks like putting on socks versus a Butler Robot AI focused on helping with more irregular events like addressing a faucet leak. Or, similarly, the indistinct gap between an Advisor AI, which is meant to serve as an auxiliary brain, and a Conductor AI, that could also be perceived as an auxiliary brain even though its domain is all about managing auxiliary muscle. Still, the preliminary framework is helpful in mapping both plausible and preferred futures for seniors aging in place through support from many different forms of artificial intelligence and robotics. It’s also helpful in revealing the human-centered design considerations we’ll need to make when developing them.
This article was originally published in Fast Company.