In brief

In brief

  • As artificial intelligence (AI) technology is maturing, the applications of AI assistants are becoming more sophisticated and transformative.
  • Zooming out, AI assistants are on track to become more like AI companions and this presents several new and interesting challenges and opportunities.
  • The article below explains the effect of this transformation, and its challenges and opportunities, with regards to elderly care.


Between 2018 and 2050, the world's total population will grow by 28 percent. At the same time, the global population over the age of 65 will grow by 126 percent.

In several countries—including Italy, South Korea, Greece, Portugal, Singapore, Japan and Spain—the population over 65 will rise from around one fifth, to over one third by 2050.

Population aging will stretch resources in communities, healthcare systems and aged care providers. So could AI help offer better support to the elderly, while also freeing the traditional resources to do the same?

AI can help improve aged care

AI shows great potential in caring for the elderly. For instance, Mabu, a portable healthcare robot, can remind people to take their medicines, maintain eye-contact and provide personalized conversation.

Accenture has also developed an AI-driven support for elderly people. It can do almost everything, from encouraging positive activities to identifying abnormalities in behavior and alerting family or friends.

AI can help with daily activities—including mobility, health and communication—while also providing “company” to those who feel isolated. This could help many countries support their aging populations, particularly by allowing them to live independently at home for longer.

AI can help with daily activities—including mobility, health and communication—while also providing “company” to those who feel isolated.

Barriers to adopting AI companions for the elderly

However, before AI-assistants can play a more integral role in caring for the elderly, there are many practical and ethical challenges and considerations that need to be addressed.

  • Digital Literacy. Those over 65 have grown up without computers, and can feel marginalized by digital substitutes, such as online services replacing their neighborhood shops or bank branches.

    On average in the U.S., only 46 percent of those over 65 have a smartphone. Internet adoption is higher, but still only 50 percent have home broadband. However, this barrier is gradually falling over time as technology adoption, even among the elderly, is on an upward trajectory.

    But an interesting point to make here is that AI can help give elderly citizens easier access to the benefits of digital technology. For example, conversational interfaces in natural language can provide better access for older citizens than touch screens and menu systems.

  • Designing company. Friendly company is perhaps one of the main benefits of AI companions for the elderly. However, AI that looks and talks like a human can be comforting to some but alienating to others. Some may be offended at being given “fake empathy” instead of human care. This raises many questions around how to design AI companions. To what extent should they emulate people and human interactions? Should they be hyper-realistic or more like science fiction?

    Having human features in AI can be reassuring, not just for elderly users, but even for factory-bound robots. Features do not need to be realistic, but in our experience, anthropomorphic designs can help exploit the cues we use to relate to each other—like smiles, speech, eyebrow raises, or gestures. That makes AI more intuitive and increases the ability of any machine to play a companionship role.

  • Funding companions. Another consideration is that many people who would benefit most from AI companions cannot afford them or would not buy them. For instance, Paro—an AI assistant in the shape of a seal that could help alleviate dementia symptoms—costs $6,400, a prohibitive price for many of those who could benefit from it.

    There’s another question: who should pay for these companions? Should people be encouraged to buy them under certain (means-tested) circumstances (e.g. for health monitoring)? Should governments fund such initiatives?

    Funding for pensions, healthcare and other types of support for the elderly is a central challenge in taking proper care of the aging population. However, rather than adding to this concern, AI is beginning to show that it could help make aged care more affordable, supporting caregivers and healthcare providers by tackling the time-and-cost-heavy support tasks.

    We have also seen that these devices can help improve wellbeing and enable the elderly to stay connected with their families and healthcare providers.
AI is beginning to show that it could help make aged care more affordable, supporting caregivers and healthcare providers by tackling the time-and-cost-heavy support tasks.

AI can increase the opportunity for human connections

The trickier question is whether it is right at all that AI, not humans, act as companions to the elderly and isolated. There are certainly many benefits to AI companions, but we do not believe AI should ever be seen as a replacement for real human interaction.

Ultimately, far from replacing humans, we hope that through AI, we can improve the quality and frequency of the connections between the elderly and their families, friends and communities.

In practice, this means that AI can support the elderly with their day-to-day needs, keeping them healthy and liberating them (and their caregivers) from some of the more time-consuming and physically challenging tasks. This could give older people, their families, friends and caregivers more time and energy to invest into quality human companionship.

Laetitia Cailleteau

Managing Director, Global Lead – Conversational AI

MORE ON THIS TOPIC


Subscription Center
Stay in the Know with Our Newsletter Stay in the Know with Our Newsletter