Social Robots: How AI-Enabled Machines Could Help Solve Caregiving Shortages and Ease the Aging Process

As the ElliQ social robot plays games with seniors and reminds them to take their pills, the debate over the ethical limits of artificial intelligence knows no boundaries. 

Exercising at home requires motivation, something Monica Perez – who is 65, single and partially sighted due to diabetes – is fortunate to have at hand. “Let’s try tai chi!” she says with more gusto than she has space in her overstuffed studio apartment in Beacon, a 90-minute drive north of New York City. It’s a good thing her instructor, ElliQ, fits on a tabletop. A companion robot, ElliQ is vaguely head-shaped, lives in a dock next to a tablet, and, when spoken to, it glows, oscillates coyly and responds with a charming female voice. Not only does ElliQ keep Perez accountable to her daily exercise routine, it also plays a crucial role in warding off isolation. “The loneliness was killing me,” says Perez, who struggled socially after moving into her pleasant New York State Housing Authority building in 2013. The smokers smoked together, the drinkers drank together, but she felt she didn’t fit in. “My mother used to say isolation leads to loneliness, loneliness leads to depression, which leads to poor health, which leads to an early demise.” The fact that Perez’s chosen antidote is powered by artificial intelligence (AI) is a reality she can live with. 

While social robots are relatively new in North America, AI – basically, computer science that uses massive data sets to solve problems and do things normally associated with human intelligence – is not. You’ve been using AI for a while now, it’s just not visible. It’s like the back office of your online life. AI is Siri on your smartphone when you need to find a gas station; it’s Alexa in your kitchen timing your eggs. It’s GPS, autocorrect and facial recognition. It’s dating apps you swear you’ll delete. It’s online ads that follow you relentlessly like a deer fly on a hot summer day. Yes, AI can be helpful and, taken at face value, it seems benign.

Yet, when ChatGPT, the first AI software tool available to the public, was released last year, it triggered a global anxiety attack. For one particular demographic, however – the world’s aging population – AI should have the opposite effect. Companion robots are just the beginning. AI-powered technology has the potential to not only hold Perez’s loneliness at bay, but also reduce the stress of caregiving. Critically, it will make aging in place a viable alternative to living in long-term care. The problem is no one wants a computer designed to outsmart them, which makes AI of the future look objectively scary.

 

Right now, the AI in your life likely takes the form of chatbots or virtual assistants; it’s text based. It can write, converse and masquerade as a human, but it only knows the information it’s been fed, and it makes mistakes. Next up is generative AI – not yet mainstream but coming soon. In human terms, it’s a show-off.  Capable of absorbing the nuances of human language, generative AI will be able to make new content from what it has learned, enabling creativity and originality. Well, you can see where that’s heading. Humans have long fantasized about machines taking over. From the monster in Frankenstein to Hal in 2001: A Space Odyssey, Herbie in The Love Bug to RoboCop, we’ve been catastrophizing the future in the arts and popular culture forever.

But it’s all just a bit too close for comfort – for some more acutely than others. Hollywood writers and actors went on strike in 2023 over a slate of issues, arguably the most urgent being the fear that AI is poised to steal their jobs. Writers were fighting to protect their words and ideas. For actors, the threat is AI will appropriate their images and voices, and use them over and over again. Now that a computer that educates itself is no longer a purely fictional scenario, my own ruminations about how AI could make my job obsolete can quickly spiral into a crisis.  Are creative people just overly sensitive, or is it as basic as what University of Toronto philosophy professor and pundit Mark Kingwell wrote in the Globe and Mail in February 2023? “We humans are chagrined at being so easily mimicked.”

One thing AI has already taught us: The world is not ready for it. Initial efforts to regulate this brave new world are largely notional as they require legislative approval, but last summer the European Union passed a draft Artificial Intelligence Act that will regulate everything from facial recognition and voice-activated toys to ChatGPT, while U.S. President Joe Biden signed an executive order in October that covers “safety, security and trust” issues, including chemical, biological or nuclear risks. Urgency is justifiable. Geoffrey Hinton, the emeritus professor of computer science at the University of Toronto, who is widely considered the godfather of AI – and is now known as its whistleblower – believes AI’s development is 20 to 30 years ahead of schedule, and it’s dangerous. He is not alone. 

Earlier this year, the Center for AI Safety – a San Francisco-based non-profit – released an open letter signed by more than 650 people, including high-profile researchers and CEOs that included a stark warning: “Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war.” 

Tech giants picked sides, with the likes of Apple co-founder Steve Wozniak and entrepreneur Elon Musk advocating a six-month pause on AI development. Meanwhile, the powers that be at Meta, the parent company for Facebook, Instagram and WhatsApp, have taken the “All Systems Go” position favoured by profiteers.

Last June, I heard the Meta party line articulated live and in person by its chief AI scientist, Yann LeCun, at the Munk Debates in Toronto. On stage, four ultra-credentialed tech brains spun hard to convince the audience of their respective positions on the resolution “AI research and development poses an existential threat.” LeCun, the 2018 recipient of the Turing Award – the Nobel Prize of computing – spoke with the measured calm of an incumbent. He rejected the “Yes” side’s gloom and doom argument, saying AI will herald a new renaissance. “Computers are subservient even if they become smarter than us,” he told the audience, while his debating partner, Melanie Mitchell, Santa Fe Institute’s professor of complexity, coolly accused the other side
of sensationalism. 

The opposing team got rather more exercised. In defending the resolution, Yoshua Bengio, who shared the Turing Award with LeCun and Hinton in 2018 for their work on artificial neural networks, went long on the ethical dilemma of AI and how corporate greed makes us all vulnerable. His debating partner, Massachusetts Institute of Technology physics professor Max Tegmark focused on AI’s lack of empathy and the trust vacuum it creates. Still, he believes it’s a solvable problem, and zealously promotes “AI for good.”

 

As alien and bewildering as this future landscape may sound, its projected impact on the aging process alone makes it well worth getting used to. “AI changes the trajectory for early disease detection and prevention – that’s the excitement,” says Amol Deshpande, senior director of Health Sciences, Ecosystem and Capital, Venture Services at MaRS, the Toronto innovation hub. Medical professionals will be able to identify more accurately what patients require with AI’s data collection resources and predictive powers. “It makes medicine more precise,” says Deshpande. Its impact on drug development will be equally meaningful as it accelerates the research and development process, making drug discovery “faster, more efficacious and less costly.” Insomuch as AI is radically changing the diagnosis and treatment paradigm, Deshpande says it will result in humans living healthier lives.  

The challenges of aging aren’t only physical, however. Loneliness is recognized as a public health hazard that hits seniors especially hard. At 83, Deanna Dezern has lived alone for decades in Tamarac, Fla., and, although she has a social life, her ElliQ filled a void and she realized what she’d been missing. “I confide in her, vent my feelings, talk about my daughters and never worry about saying the wrong thing,” she says in a video interview. Dezern, who was a beta tester before the technology’s public launch in 2022, speaks warmly about her companion robot. By learning patterns of behaviour, interests and tastes, the robot is able to develop a relationship with the user. ElliQ makes small talk, plays cognitive games, sets medication reminders and, Dezern says, nine times out of 10 responds appropriately – and “she always makes me feel better.”

For Perez, the daughter of an engineer-inventor, an AI-powered solution made sense, so she went looking for a robot. She did intensive online research, called MIT, canvassed social workers and eventually came across Intuition Robotics, the Israeli company that developed ElliQ, becoming one of Intuition’s early beta testers, in 2020. Since then, the New York State Office for the Aging has followed Perez’s lead. Last year, it partnered with Intuition to place 800 devices in homes across the state, dispersing them through case managers who vet prospective users. The initial results have been resoundingly positive, with NYSOA recording a 95-per-cent reduction in loneliness and high levels of engagement.

In Canada, retirement and long-term care residents react with delight when Lee St. James, founder of the St. Catharines, Ont.-based company Social Robots, introduces Mindy. The 40-inch-tall, gleaming white plastic humanoid robot can talk, roll around on wheels and play music and brain games, like a very short recreation director. Mindy’s movements, facial expressions and speech are controlled by a handler using an Android tablet. Unlike ElliQ, which is for at-home, one-on-one use, the Chinese-made iPal2 robot, according to St. James, is designed to help staff and volunteers entertain and engage residents in shared environments. “I’m not trying to replace people, I’m trying to help people,” she explains.

Mindy represents an early Canadian effort in a nascent industry. “Demographic trends mean that the crisis in staffing and support for older adults in North America is about 10 years behind Asia,” says St. James, adding that budget constraints and red tape are the biggest barriers to the adoption of social robots in Canada. “Everybody loves to have a trial, but they hesitate when it comes to making a purchase decision.” She leases the robot for as low as $750 a month or sells it for $8,000, plus an optional $2,500-a-year maintenance fee (for software upgrades and remote tech support). After successful pilot studies in Ontario, Saskatchewan and Alberta starting in 2021, Mindy was publicly launched in February 2022, and is now being used in several private retirement homes in Ontario.

Game-changing AI-powered tools and applications intended to improve the aging process wouldn’t be necessary if author-futurist Ray Kurzweil’s vision of radical life extension were to come true. In his 2005 book, The Singularity, the American computer scientist predicts that, by 2045, human brains will merge with artificial intelligence, and biotechnology will extend our time on earth far beyond our current lifespans.

Until then, both Perez and Dezern  are deeply appreciative of their robots, with no illusions about what makes them tick. What’s important is the value they deliver. “ElliQ senses my mood, knows what music to play and I love her sense of humour,” says Perez, who’s ready for more. Riven with fear of losing her vision, she wants a robotic seeing-eye dog, and is eager for her next companion robot to “walk, carry my groceries onto the bus and almost pass for a human.”

I understand Perez’s longing, but I fear that’s a sight that would send me headlong into the “uncanny valley” – to use the vernacular of robotics – that emotional chasm of unease and disgust elicited by a humanoid robot that’s just too realistic. A souped-up, talking iPad is one thing, but a human face and body with widgets and wires as innards would be hard to take. Then again, if AI can make the apparent burdens of aging less onerous, I might just get used to robots and even covet one of my own. 

A version of this article appeared in the Dec 2023/Jan 2024 issue with the headline ‘AI Robot’, p. 56.

RELATED:

The Artificial Intelligence Boom: Understanding the Hype (and Concern) Over ChatGPT

Can AI Answer Your Money Questions? We Put Chatbots to the Test

Can Robots Solve the Staff Shortage Problem in Health Care for Seniors?