For many of us, artificial intelligence has already become part of our lives: every time we ask a smart speaker to give us a news briefing, or play us a favourite tune, we are engaging with AI.
But today’s thinkers are also looking to AI for solutions to some of the major challenges facing society, like the ageing population. It’s estimated that by 2050, the rich world’s population, including the UK, will have less than two people in the 20-64 working age bracket for everyone over 65. That’s a massive shift from the current position, where there are around three people of working age for everyone over 65. This shift will create even more pressure for states and for families.
Friend or foe?
AI, so its champions believe, has the potential to help people age better, maintain their independence for longer, even combat loneliness. But there are others who fear the negative consequences of the so-called rise of the machines, with even some tech bosses such as Google’s Sundar Pichai warning that AI is a threat to humanity.
These debates are nothing new to Jacques Fleuriot, who holds a personal chair in artificial intelligence at the University of Edinburgh. “What a lot of people don’t realise is that AI – in one form or another – has been around for quite a while, but it’s only now it’s really properly hit society and is in people’s consciousness,” he says.
“If you look at the history of AI, it’s definitely had its ups and downs in terms of being flavour of the month or not. But for me as a young student, I just thought at the time it was something that looked exciting, without fully appreciating the way things would be by now.”
The next generation
Europe’s first AI research hub was forged at the University of Edinburgh 60 years ago. Today the University remains at the forefront of research as well as teaching and bringing on the next generation. Finding ways of using AI for better ageing is a key element of that.
Prof Fleuriot is a leading figure at the University’s Advanced Care Research Centre (ACRC), a multi-disciplinary programme combining medicine, engineering, informatics, and social sciences, which aims to support older people to live with dignity, independence, and a good quality of life, in their own homes and in supported care environments.
It’s not just about the technology, says Prof Fleuriot, it’s also important to ensure it is acceptable to those who will be using it. That means, for example, developing unobtrusive sensors that can be incorporated into someone’s home or environment in an ambient fashion, rather than something they have to wear next to the skin, or charge every night. “Wearable devices are great, but don’t necessarily work for a particular target population for multiple reasons. So we’re developing different types of sensors that can detect physiological signals – for example, a ‘smart’ fabric that can measure the level of hydration in a person who is sitting on it.”
End-user involvement
This sounds simple, and benign. Who could argue with a device doing something to keep people safe and well and independent? But it is fraught with concerns about privacy, for example, and when monitoring someone and sharing their data goes beyond what someone would accept. That’s why it’s so important to involve potential end-users of a technology at every stage, says Prof Fleuriot.
“I think ethics is one of the biggest problems we have, and I’d be lying if I said we had the answers to all of this. Because as the technology is evolving and the more you can do with it, part of the quest is to answer some of these ethical questions. Designing it with people is so important because you can understand about what is acceptable to them, but also if you address the ethics early on. It’s baked in.”
Ideally, the process of helping someone remain independent for longer starts well before they start to exhibit signs and symptoms of deteriorating health. That’s so the technology can understand a person’s baseline in order to detect changes that might indicate increasing frailty, for example.
Jane Hillston, professor of quantitative modelling, explains: “One of the things AI can do, and what we’re aiming to do, is extending people’s quality independent living by developing devices and modelling techniques that will allow unobtrusive monitoring in people’s homes to make sure they are active and carrying out their daily activities,” she says. “But what we’re also doing is assessing for each individual ‘what is their normal?’. How long do they typically take to climb the stairs, how often do they drink tea or go to the bathroom? And you can have models that allow you to detect if something is normal for a person, or, as you collect the data, you can say ‘okay, something’s deteriorating here’.”
Enriching your environment
AI can also be used more obtrusively, she adds – for example, cameras might show that someone has fallen, and an AI device might ask the person if they are okay, if they can get up, and reassure them help is coming.
Seemingly small things can make a big difference, she adds, such as cutlery that can tell if someone is developing a tremor that could indicate Parkinson’s disease. “I like to think of it as enriching your environment – and helping you to live independently for longer.”
The pace of change is rapid. “It’s creating new opportunities,” says Prof Hillston. “In some ways we almost need to speed up our imaginations to see the best ways of taking advantage of advances in technology. But they do need regulation. Like any new tool, it can be used for good or bad, and we’re not there yet either in setting social norms or setting the kind of legal and regulatory and compliance infrastructure to ensure we’re maximising the positive use and minimising the negative use.”
Tackling loneliness
AI also has the potential to help people feel less lonely and isolated. According to the charity Age UK, 1.4 million older people in the UK are often lonely, with a serious impact on their health and wellbeing. Again, Edinburgh researchers are working to combat this.
“We’ve got a very large group working on natural language processing – chat bots and similar technology,” says Prof Hillston. “I think there is great potential in making people feel less isolated.”
‘Smart pets’ are one example. She describes a collaboration between University of Edinburgh design informatics students, and robotics students from Heriot Watt which resulted in a sea-anemone-type toy made from a knitted fabric that responds to touch. “You touch them and they sort of arch their back,” she says. “It’s not threatening to anyone, it doesn’t have a face, it’s not pretending to be anything but an object, yet it’s a responsive object that you can stroke and it will physically respond. I think my mother, who is something of a technophobe, would like something like that.”
Intelligent robots
Ram Ramamoorthy holds the personal chair of robot learning and autonomy in the University’s School of Informatics. He describes his work as being at the intersection between robotics – devices that physically interact in the world, such as humanoid robots or autonomous cars – and machine learning. “We’re interested in how machines learn to improve their models in a data-driven way, and adapt their policies for action,” he says. “We want to make machines that can, to some extent, act by themselves in human-centred environments.”
His long-term vision is of care robots that, rather than replacing humans, are intelligent devices that augment people’s capabilities or provide abilities they are missing, perhaps as a result of neurodegenerative diseases. He gives the example of someone who previously loved to cook, but who has developed Parkinson’s disease or dementia, and experiences cognitive problems, such as forgetting how to use kitchen tools, or physical symptoms such as a tremor, which can make using the tools difficult.
“This all translates to a quality of life loss for someone who used to be able to do these things and pride themselves on being a good cook,” he says. “So we are interested not so much in bringing ‘Rosie the Care Robot’ into someone’s house, but to make parts of what they use – in the spirit of smarter tools and assistive devices – more intelligent.”
Supporting independence
This could involve robotic devices that help compensate for the loss of capacity, in a similar way to using a prosthetic device to compensate for a lost limb, he adds. “Food prep is a good example, and right now in the robotics lab we study tasks such as manipulation of soft materials – we can peel oranges and cut vegetables. We can do that with general purpose robotic systems we have in our labs. The next question is, what would it take to package this in customised devices that can go into someone’s home and do the job for them, while still allowing them to be in charge at a high level? That form of shared autonomy is a fairly difficult technical challenge.”
Shared autonomy is important to maintaining independence, he says, because the more actively engaged a person remains the better for their overall health and wellbeing. “The more they switch off their brain, the worse their brain gets,” he says simply. “And the more they don’t use a capability, the more it deteriorates.”
At the heart of the cross-sectoral and multi-disciplinary work is a wish to help people to age better – and he believes AI can help. “People want their current lives to be maintained and improved when they’re losing capacity, so they want to fill in the gaps,” he says. “But these are all complex problems covering all facets of life, so there’s going to be no one solution.
“At Edinburgh we’re a diverse community, which is how it should be. And we’re all learning to make the best use of each other’s expertise.”
Picture credits: older woman with robot – wonry/Getty; Sundar Pichai – Justin Sullivan/Getty; Maah companion robot – Konpanion; older woman cooking – Jean-philippe WALLET/Getty.
