Episode Transcript

[00:00:00] Zannat Reza: Artificial intelligence is being talked about a lot these days. Also known as AI, it's created a buzz with chatGPT bursting onto the scene. Now, chatGPT is a tool that harvests information from across the web to answer your questions by generating conversation like text. But there's also been public debate about how this tool might impact the future. Things like privacy issues, potential job loss, and whether there needs to be regulations. Is it all hype? Or is there real potential to accelerate the work we do? And crucially, how best can AI be harnessed when it comes to the future of aging?

[00:00:34] I'm Zannat Reza. Welcome to The Future Age Podcast, where we explore creative solutions in re-imagining what life could look like as we get older.

[00:00:42] If you enjoy this episode, please follow or subscribe on your favorite listening app. For this episode, we're exploring the potential and pitfalls of AI in the aging space. I asked chatGPT about this and it wrote, "AI can be used to improve the quality of life for seniors. AI can be used to provide [00:01:00] remote monitoring, detect cognitive decline, manage medication, and provide social and entertainment activities."

[00:01:06] Well, thanks ChatGPT, that's promising, but let's unpack what's possible in a practical way. We have two special guests who bring a global and pragmatic lens to the world of AI. First up is Dr. Charlene Chu, an assistant professor at the Faculty of Nursing at the University of Toronto. She's also cross appointed with the Institute for Life Course in Aging and is an assistant editor at the Canadian Journal on Aging. [00:01:29] Her research includes how artificial intelligence can support healthy aging and aging in place, and co-designing tech solutions with older adults. So, Charlene, what exactly is AI?

[00:01:43] Charlene Chu: Artificial intelligence, otherwise known as AI, refers to the ability of machines to perform tasks that typically humans can do, and that require human intelligence. So, being able to understand natural language, recognizing objects and [00:02:00] images, making decisions based on information that is being provided. For example, Alexa or Siri, these are AI powered systems where you can speak into them, they understand your voice, and it is understanding sound as language and taking that audio and following those commands.

[00:02:20] Zannat Reza: What are the opportunities for AI in the aging space?

[00:02:25] Charlene Chu: AI is going to be revolutionary in aging, as well as in health, it has the potential to improve the quality of life of the aging population in a variety of ways. AI can be used to improve the quality of care that can be delivered to older adults who are living at home. So, for example, we can have AI that can help predict conditions or diseases for people who are living at home.

[00:02:50] The same information could also be used, for example, in the hospital. We could use AI to predict hospitalization outcomes for older adults before they go home. We can look at recovery [00:03:00] trajectories. It can be used as a very powerful tool in the prediction of different conditions that impact older adults. Smart home technology to help enable older adults to live independently at home; that can help with medication management, home management. When we think about smart home technologies, we can also imagine technology that is detecting falls and then sending an alert to a caregiver, whether that be a family member or a clinician.

[00:03:26] Zannat Reza: So, for medication management, how can AI help in that specific task?

[00:03:32] Charlene Chu: Being able to provide medications on time, keeping track of the different medications that an older adult might take, from a caregiver perspective, is a big contributor to caregiver burden, so we could have medication reminders. Different educational tools as well, that can be placed around the home, to remind the older adults to take medications or the caregiver to take medications.

[00:03:58] Zannat Reza: Charlene goes on to tell me that for apps that manage care delivery, pharmacists can monitor and communicate with an older adult at home and adjust medications as needed. But not everything is rosy when it comes to AI for older adults. Charlene points out that digital ageism or age discrimination is a big issue in these technologies. These negative stereotypes and attitudes towards older adults can creep in and influence how we design, implement, and evaluate technology.

[00:04:24] I asked Charlene to explain how this plays out.

[00:04:27] Charlene Chu: Research has shown that even when older adults are there in front of someone designing the technology saying, this is not what I need, this actually is not helpful to me. The designers will continue on down the path that they originally intended because of their own beliefs of what they think in older adult needs.

[00:04:44] When you look at it from a data perspective, when we're talking about AI and machine learning that is data driven, we look at who is able to even participate in having their data collected. Often the people who are able to participate in having their [00:05:00] data collected are younger people, university students, people who can have access to various types of technology that may be quite expensive, people who have access to internet.

[00:05:12] Zannat Reza: These biases even extend to a technology that uses facial recognition.

[00:05:17] Charlene Chu: These data sets are incredibly commonly used, and they contain tens or hundreds of thousands of pictures of people in order to train models to detect people's faces. We found that there was less than 0.05%, less than half a percent, of pictures that were representing older adults.

[00:05:39] You can think about how you turn on your computer, it probably uses facial recognition. How you open your phone, it probably uses facial recognition. So, all of these types of either the presence or absence of data in these larger data sets, can influence the accuracy of the algorithm. It can impact who is able to benefit from the algorithm, and who's not able to benefit from the algorithm.

[00:06:03] Zannat Reza: Age biases can even influence job searches and the kind of apps that are available in the app store.

[00:06:09] Charlene Chu: Many social media companies, um, as well as companies that are looking to hire, will often only target people who are younger. And so, they're able to do that because, if you are an older adult and you might be looking for a job in a specific sector, or you are of a certain age, you know, [00:06:30] all of our data is being collected. They know how old you are, they know your, your gender, um, you know, they know your location, and so they're able to identify somebody of an older age, and they may not advertise that job to you. If we look at apps in the app store for older adults, that the vast majority of them are about medical care and managing chronic conditions, that's what we think older adults are interested in.

[00:06:57] Even if we think about games for older adults, you and I have lots of choices about the kinds of games that we want to play using our iPhones, uh, using our Android phones, right? So, there's tons of different games that you can play out there, whether it be Candy Crush or whatnot. But when you look at games for older adults, it's about preventing dementia and cognitive stimulation. It's not just about, here is a game that you might enjoy, that you might like to play. It is targeted towards having some type of health benefit because we think that older adults have this biological impairment.

[00:07:33] Zannat Reza: So, if I understand correctly, what you're saying, is a lot of the apps that are developed for older adults really just focus in on medical conditions, whereas the developers may not be thinking, oh, older adults may also want to play games that are not health focused. And actually, I'd be curious to know how many older adults play Candy Crush.

[00:07:53] Charlene Chu: Would be a great study.

[00:07:55] Zannat Reza: How do we make sure that AI is inclusive?

[00:07:58] Charlene Chu: I think involving older adults in the design process, um, making sure that their needs are taken into consideration is really important. So, this can look like doing user-centered research and co-creating prototypes, for example, soliciting feedback as prototypes that are being created, having consistent testing, iterative testing.

[00:08:23] Zannat Reza: While creating technologies that actually meet the needs of older adults is important, many of these solutions collect data, which leads to a bigger discussion around data privacy. A prime example here is smart home technologies that allow people to live independently in their home.

[00:08:38] Charlene Chu: When we think about smart home technologies, privacy is a huge issue here because it might collect quite sensitive data about older adults living in their home. There's a big question around what they would want to be collected versus what they don't want to be collected. And, we end up in this kind of situation where we may be faced [00:09:00] with a phenomenon that's called digital resignation, where when you're provided a choice that whether to engage or not with a different type of technology, that you end up engaging, because you are digitally resigning yourself.

[00:09:15] Many older adults might not be so into the idea of having their data collected from knowing when they turn on their faucets, how many times they go to the washroom, how many times they leave their house, you know, movement in their [00:09:30] bed, how many times they open the fridge. An example of that might be, you don't want to sign up for Facebook, but all of your friends are on Facebook, and that's the only way that you're going to see pictures of your niece or your nephew, so you sign up for Facebook.

[00:09:42] A really sensitive topic for older adults is, well, what happens when I can't live at home? In that case, maybe I don't want other people to know that, because of safety concerns, and that might mean that I need to move into a long-term care home, and I don't want to live in a long-term care home.

[00:09:58] But from a healthcare perspective, that's kind of important for me to know that you can open your fridge and that you're able to access food, and that you're able to access food on a consistent time period throughout the day, and that you're able to eat your meals. And so that is an indicator of whether or not you're able to take care of yourself and live independently.

[00:10:17] And if they're not happy with what's being collected, how do they get that data back? I'm not sure that they can, and they also may not have a say in who gets to see that data. So, is it their family member that they may, or may not, have a good relationship with? Is it a family provider that they may, or may not trust?

[00:10:37] Zannat Reza: Data privacy is one issue, but another big question is who owns that data? For that, and examples of AI and aging from a global perspective, I spoke with Stephen Johnston, founder of Fordcastle, a company that works with clients from around the world to tackle society's biggest challenges and opportunities, including aging and longevity.

[00:10:56] We tracked down Stephen in Australia, and ironically, for an episode on AI, the tech was a little glitchy. I asked him what types of issues stem from collecting AI data?

[00:11:05] Stephen Johnston: Oh, a ton, a huge amount of data of issues stem from it, and I think that's a really important conversation that's happening now. It's been a little late, I think the tech companies, the large tech companies, have made billions and 20 trillions off the lack of anybody asking those questions up until now, about who is using that data. And they've been showing ads, [00:11:30] and we are receiving ads, and the tech companies are being paid for the privilege of us spending our time watching those ads, and that's obviously backwards. I mean, if we're seeing ads, we should be benefiting. We should be the ones to say, "You know what, I will allow you to have some of my attention, and some of my time, and potential buying power" and that's not the case.

[00:11:51] Where does the data come from? Who owns the data? How transparent is it, and how is it actually delivering the results and those insights that you're [00:12:00] potentially, you know, then sharing back. But there needs to be much more transparency about the process of who owns the data, and then how is it used to deliver those insights. I think effectively, the right answer is, you as an individual own it and you own your contributions, and the reality is, that it's quite hard in practice to make large language models work, and AI work, without having sort of, massive amounts of data that has been generated from vast amounts of people.

[00:12:29] Zannat Reza: Owning your data is one thing, but what about being able to benefit from your data? After all, it's yours, and if you own the data, shouldn't you be able to decide how it's used and be rewarded for those decisions?

[00:12:41] Stephen Johnston: Not that many people have taken ideas like this to scale. I think they're really hard and complicated to do, but it's the right question to be asking. Who's benefiting from the data and who's owning it, and more importantly, how can you ensure that you can control the data? Because ownership is one thing, and control is the other.

[00:12:59] Zannat Reza: So even though data ownership and issues around use need to be ironed out, there are some great examples of AI solutions in the aging space that exist right now.

[00:13:09] Stephen Johnston: AI for caregiving has real potential to deliver the job of what needs to be done, in high quality care, in the right time, in the right place without focusing on staffing ratios. Because there could be ways, and Sampo, one of my clients in Japan, is a big insurance company, but also runs the most nursing home beds in Japan. [00:13:30] They were able to shift the ratio of caregivers from one caregiver for three residents, to one caregiver for four residents using AI, using a combination of tech and data to provide better care and better outcomes, while also reducing the number of caregivers in a company.

[00:13:49] TapRoot is using AI to look at questions and responses that people have based on certain situations, and so the expectation here is that you're a caregiver [00:14:00] and you don't really know what to do with somebody, for example, they may have dementia. The AI model would allow you to be given prompts, and to train it, based on whether the prompts and suggestions of what to do, is effective or not. And so, if it turns out that this kind of intervention works well, then you say it's working well, then other people can benefit and learn from that. And over time, we're building up, essentially, a way to have those without much of a caregiving skill or [00:14:30] expertise, given access to the real cutting-edge insights about how best to treat, care for, and connect with, people from different walks of life and different personas in any, uh, situation that arises.

[00:14:44] Zannat Reza: Stephen goes on to say that AI takes large amounts of data, looks for patterns, and makes sense of what it all means. Companies like Care Daily and care.coach are making it easier for caregivers, health professionals, and older adults to communicate and coordinate services.[00:15:00]

[00:15:00] Care Daily's AI powered app collects information from sensors, connected devices and online information to establish baseline data so it can identify when things veer out of normal boundaries. It can identify in real time when people fall and can alert caregivers or emergency services. It can also uncover health problems.

[00:15:19] As for care.coach, it has several AI options. Its Avatar Service can act as a social companion and caregiver by doing health check-ins, coaching clients to take care of [00:15:30] themselves, and offering medication and appointment reminders, among other features. And a Canadian company, Winter Light Labs, has developed an app that you can speak into and the AI is able to pick up over 400 linguistic cues undetectable to the human ear that can be used to gauge the likelihood that someone may develop dementia or a mental illness. I mean, how cool is that?

[00:15:53] And now I wanted to find out what Charlene and Stephen think about the future, including their own. I ask them the two questions we ask every guest. First up, finish the sentence in 10 words or less. The future of aging should be....

[00:16:06] Stephen Johnston: The future of aging should be the future of society. Meaning there's no difference between aging and society.

[00:16:15] Charlene Chu: I think the future of aging should be dignified, inclusive, and empowering.

[00:16:21] Zannat Reza: Now let's time travel to when you're a hundred years old, what does your ideal life look like?

[00:16:27] Stephen Johnston: I will still be working; I will still be doing podcasts. I will be, hopefully, by that stage, with more balance between work and life, and I think I'll be on a gradient. I'll be more life, personal, family, than work, work, work, which I'm currently right now.

[00:16:46] Charlene Chu: At the age of 100 my ideal life would involve living independently, with access to community resources and a supportive network of family and friends who love [00:17:00] and care for me. I would be physically active, mentally engaged, um, pursuing activities that bring me joy and fulfillment.

[00:17:08] I would have access to healthcare and support services that continue to help me live independently and have a high quality of life as I age. And I think I would be living in a society that would value the wisdom and experience that comes from age.

[00:17:27] Zannat Reza: A big thank you to Stephen and Charlene for a thought-provoking discussion on the potentials and pitfalls of AI. And thanks for joining us for this episode.

[00:17:35] To learn more and for transcripts, go to thefutureage.ca. Listen to new episodes by following us wherever you get your podcast. And if you're liking our podcast, leave a review on Apple or Spotify and be sure to share it with your friends, family, and colleagues.

[00:17:50] The Future Age is brought to you by SE Health, a not-for-profit social enterprise, whose purpose is to bring hope and happiness to the lives of Canadians. [00:18:00] It's produced by the Future of Aging team and Podium Podcast Company. For more information, visit thefutureage.ca.