AI-powered assistants could be the key to providing care for an aging population.
The world’s population is aging and this dramatic increase in the elderly is proving to be a challenge to current methods of care which rely strongly on in-person monitoring. This situation is only likely to intensify over the coming decades as life expectancy continues to increase.
The US census bureau¹ estimates that the global population of individuals over 65 will grow to 17% by 2050, roughly equivalent to 1.6 billion people.
Longer life spans will not just result in an increase in the population of older adults, however. Increased life expectancy will also result in an increase in people living with physical impairment and chronic conditions that impact their ability to perform daily activities.
With the demand for elderly care growing, and a shortage of direct care workers in many countries, researchers are increasingly developing artificial intelligence (AI) and machine learning tools to assist in the care of the aging.
Increasingly, health monitoring tech is playing a role that would have once been fulfilled by human care providers caring for older adults and helping complement care provisions. Whilst this may initially sound somewhat cold, the use of AI is, in many cases, allowing elderly relatives to care for themselves and stay in both their own and family homes for longer. Ultimately AI is helping the elderly maintain their independence for longer.
Can AI Remotely Care for an Aging Population?
One of the most straightforward approaches to boosting caregiving with AI is by using technology to monitor health and well-being remotely. Whilst video cameras can observe older adults in their homes, helping them maintain independence, these systems still rely on human operators.
As well as being labor-intensive, this means these systems are also vulnerable to human flaws such as distraction and just old-fashioned human error. That means that an AI health monitoring system could actually offer advantages over human-operated systems².
And these systems can also be less intrusive. Many of the apps we use on our smartphones are already monitoring health data, including activity, diet, and even snapshots of our lifestyle in the background of our daily lives. AI could also monitor more ‘in-depth’ health data with blood pressure or electrocardiogram monitors may help to predict various health conditions such as hypertension and atrial fibrillation.
But this isn’t the extent of the task that an AI care assistant could provide for older adults. Sophisticated machine learning can even learn to classify activities like standing and walking, knowing what is usual for a particular older adult in a particular setting like their own home³. These models are becoming so precise that they can even identify changes in movement that could indicate cognitive decline.
Monitoring with AI can assist in the care of older adults outside the home too, using GPS to track the movements of older people. If this is considered an intrusion into privacy, monitors can simply be set to register how long an older person has been out of the home without returning alerting caregivers of prolonged absences.
This data could be used by machine learning programs to build a picture of a typical day in the life of an older individual, alerting relatives and health care providers if these activities deviate significantly. In less extreme cases such programs could also make suggestions for care provision and assess any potential health risks.
Of course, diagnostics, logistics and machine learning are all fine and good, but there is much more to care provision than this. Even if AI can deliver benefits for the medical care of the elderly, can it deliver the same human touch needed when caring for a human being?
Is There a Softer Side to AI Care?
Critics of the creeping influence of AI in the care of older adults are likely to highlight the fact that machine learning is no replacement for the relationship that develops between an individual and their doctors and carers.
And there is a practical advantage to such a relationship. For example, a human doctor may treat a person rather than just a condition by both learning about their patient, forming a bond with them, and reading unconscious cues to provide more individualized health care.
In 2019 a team of researchers from the University of Rochester set about investigating the relationship between older adults and a digital avatar⁴.
The researchers found that subjects in the test who interacted with the chatbot over a period of weeks became more comfortable with sharing more than just superficial details and details about everyday tasks with the AI. They also became more comfortable in divulging details about intimate topics such as health.
AI can also be used in chatbots to help encourage elder citizens to take part in regular self-care routines and even encourage social activities, letting them know when events are occurring locally and prompting them to go out and interact with others.
Social robots can help users take videos and photos, arrange telephone and video calls to friends and family, and even provide games and activities for younger relatives to encourage visits. All of this ultimately reduces isolation and loneliness, making an AI-powered social robot or a chatbot a source of companionship, comfort and security.
Developments like this have led experts such as former Google CEO Eric Schmidt to predict that one day AI chatbots will provide companionship for around 40% of elderly people, helping battle loneliness — something that has been shown to increase the likelihood of clinical dementia and even mortality.
Are We Ready to Hand the Care of Older Adults to AI?
Of course, even with all of these potential benefits, there are still concerns about care provisions for older adults provided by AI.
The accuracy of such systems must be robustly assessed and there are considerable privacy concerns that can be raised by the use of any monitoring system whether powered by a human or an AI.
Anita Ho is the author of a paper published in the Springer family of journals that reviews the provision of care by AI. She points out that such measures raise significant ethical concerns such as the effect on interpersonal relationships that need to be fully and frankly assessed as this technology emerges.
The fact of the matter is, robotic assistance in the care of the older generation began some time ago. Life alert pendants and nanny cams have been employed by families and care agencies for many years. Other monitoring sensors are now becoming more and more common.
The developers of such systems are reporting that their technology is regularly improving the lives of elderly users.
As an example; Safely You⁵ is a Bay Area company that employs an AI system to help with fall prevention in its memory care facilities. Its executive director Sylvia Chu says that hospital trips have reduced by around 80% since the system, which is trained to detect falls, was installed across its 23 facilities.
Only time will tell if AI systems are the answer to elderly care. At the moment though, it looks probable that such measures will only provide a helping help for caregivers for the foreseeable future.
1. He.W., Goodkind. D., Kowal. P, US census bureau, , ‘International population reports, P95/16–1, an aging world,’ US Government Publishing Office.
2. Ho. A., , ‘Are we ready for artificial intelligence health monitoring in elder care?’ BMC Geriatrics, [https://bmcgeriatr.biomedcentral.com/articles/10.1186/s12877-020-01764-9]
3. Luo. Z., Hsieh J. T., Balachandar. N., Yeung. S., et al, , ‘Computer Vision-based Descriptive Analytics of Seniors’ Daily Activities for Long-term Health Monitoring,’ Machine Learning Healthcare, [2018;85:1–18.]
4. Razavi. S. Z., Schubert. L. K., Van Orden. K. A., et al, , ‘Discourse Behavior of Older Adults Interacting With a Dialogue Agent Competent in Multiple Topics,’ [https://arxiv.org/pdf/1907.06279.pdf]
5. Safely You [https://www.safely-you.com/]