Can robots make us less lonely?
Last year, the University of Michigan’s National Poll on Healthy Aging found that 1 in 3 adults between 50 and 80 years old said they felt isolated. Not surprisingly, people who live by themselves are more likely to feel lonely.
Enter ElliQ, the robot companion made by Intuition Robotics. It was created to alleviate loneliness in older adults.
At about a foot tall, she — and the company did designate the machine a “she” — is designed to sit on an end table in the home. She’s programmed to be inquisitive, empathetic and even silly.
Marketplace’s Lily Jamali spoke with Dor Skuler, CEO and co-founder of Intuition Robotics, about why he thinks a robot is the right tool to address loneliness. The following is an edited transcript of their conversation.
Dor Skuler: Last year, the surgeon general of the United States issued an advisory identifying loneliness and social isolation as a public health crisis and attributing the health risks as equivalent to smoking a pack of cigarettes a day, and actually contributing more deaths than anything else in the U.S. right now. More than obesity, more than drug abuse. So, it’s a really, really big issue.
Lily Jamali: And it can have real physical health consequences.
Skuler: Correct. People who are lonely and isolated have double the risk of dementia, heart disease, depression and stroke. And it costs the health care system a significant amount of money. It also costs the economy because caregivers, we all become caregivers and we lose working days, and that’s an estimated $140 billion a year of lost productivity. So it’s a big, big issue.
But the numbers aside, it’s a deeply human issue. We as humans are not meant to be alone. And yet many of our elders are by themselves almost every day, all day, especially after losing a spouse. At Intuition Robotics, we believe that technology can help. It’s not a replacement for human contact but, keeping in mind that the situation is that people spend a significant amount of their time by themselves and we can’t just snap our fingers and make people show up, it does have the ability to be very, very effective. In fact, we’re seeing data from the state of New York, which issued data after deploying hundreds of ElliQs for over a year, that 95% of the people living with ElliQ are seeing a reduction in their loneliness and social isolation scores. And also 96% of them are seeing an improvement in their health and wellness indicators.
Jamali: So how does ElliQ deliver results like that?
Skuler: It starts with thinking differently about an AI. It’s not a command-and-control machine that you give a prompt like what’s the weather and it answers, although you can, of course. It’s more of a companion and a friend which is empathetic. So first off, she’s proactive, she’s not reactive. Though ElliQ will initiate interactions, she will do it in a way that’s deeply contextual and goal-oriented. She’ll remember previous conversations, and she’s trying to help you achieve specific goals around healthy aging. She remembers those things, she puts them in context. She offers interventions to help you deal with those things, whether it’s interventions she can do by herself or interventions with the outside world. Like, “Can I tell your daughter about this? Maybe she can help. Do I have your permission to tell your doctor about this? He should know.”
Jamali: I’m struck by the fact that she asks for permission. This can be very important to older people who are not necessarily wanting to always share personal information.
Skuler: It’s a superimportant point. ElliQ will not be successful if she doesn’t have the trust of the older adults she lives with. They interact with her a lot, over 30 times a day even after six months. To retain that trust, you have to decide very early on whose side you are on. And that’s even harder when the people paying for ElliQ are the government services, the elder affairs or health care providers. So, they have a certain interest when they provide them, and we had to decide whose side we are on. Usually, everyone’s incentives and objectives are aligned, but not always.
We decided that the older adult is the person we care for. The older adult is the person that has agency. And therefore, if you’re not feeling well, and you talk to ElliQ about it and she feels like it would be helpful if your doctor knew about it because they can help you deal with it, she will not let them know unless you give specific approval. And we won’t do it once in the terms and conditions in paragraph 17. We will ask every single time before we share that information to give the older adult agency over their own data.
Jamali: So how does ElliQ 3.0 differ from previous versions?
Skuler: There are two main differences. One is the hardware. It’s built for a lot more scale as we’re seeing more demand, and it obviously has a faster processor and a better screen.
But the real kind of news is also about the software. We’ve incorporated large language models and transformer models into this product now, which allows us to have much deeper conversations with individuals, which are not limited to what we imagine people will talk about with her. In the past we had to kind of try to anticipate where conversations would go and create dialogue around those areas. Now, with large language models, we can basically talk about any topic without losing what makes ElliQ superunique, which is the ability to be proactive and contextual. She remembers what happened in the past and understands that not all memories are relevant for every conversation. You need to understand the time decay of memories. For example, if you didn’t feel well last week, you wouldn’t use that in conversation the same way you would if you didn’t feel well a year ago. But, if you lost your spouse a year ago, it’s still very relevant, right? So, it’s very contextual understanding of the time relevance of each memory. And of course, putting that into special character of ElliQ and putting guardrails around safety and privacy around that.
Jamali: Can I ask how you got interested in age tech?
Skuler: I used to be a senior VP at a telecom company, and I was flying all over the world. And my daughter, who was 3½ at the time, grabbed me on my way to the airport one day and asked me, “Dad, is it important what you’re doing that you’re never home?” And that question stuck with me for this 15-hour flight, and I felt it wasn’t important enough. And I really wanted my next company to be — this is startup No. 5 for me — to be important. I don’t know if it will succeed or not, but at least if it will succeed, it will be impactful in people’s lives.
I think there are not enough innovators in this field. There are not enough products that are delightful and that really help the older adult as opposed to being geriatric and trying to monitor the older adult. It’s a very, very big difference. And when you just look at the demographics, the population is getting older and lives much longer, and there are not enough young people that are willing and able to spend significant time with them. We need that company and we need social interaction because we’re human. So, I thought that’s a problem worth trying to solve. And knowing that it’s a very, very tough problem because to solve it effectively, you need to create a digital friend and not just a digital utility.
As Dor Skuler pointed out, an artificial intelligence-based companion like ElliQ can be a tool to help keep older adults engaged, but it doesn’t replace human contact.
But even using ElliQ as a tool has some ethicists concerned.
One researcher who spoke to The Washington Post about ElliQ last year worried the chatbot’s presence could lead to a greater sense of loss if it suddenly goes away. She also noted that older adults may not fully understand that they’re speaking to an AI device.
As The Post reports, ElliQ’s friendly demeanor gets at the heart of what satisfies us about friendship, that another person cares for us and remembers our likes and dislikes, trips we’ve taken and memories we cherish. But in the case of ElliQ, the chatbot remembers those things not because it understands them, but because they’re data points that can be used to program future interactions.
When I met ElliQ, I found her to be a lot more personable than I was expecting for a robot. She’s also clever. When we asked her to create an image of two people listening to NPR, she came up with one in seconds and displayed it on her touch screen. The cartoon image featured a superhip guy and girl, with what appeared to be a latte in the foreground.
Stories You Might Like
The future of this podcast starts with you.
Every day, the “Marketplace Tech” team demystifies the digital economy with stories that explore more than just Big Tech. We’re committed to covering topics that matter to you and the world around us, diving deep into how technology intersects with climate change, inequity, and disinformation.
As part of a nonprofit newsroom, we’re counting on listeners like you to keep this public service paywall-free and available to all.
Support “Marketplace Tech” in any amount today and become a partner in our mission.
