Hey Siri, How Do I Feel?
Artificial Intelligence (AI) is playing a bigger role in patient care, in mental health diagnoses, and in analyzing big data to deliver more targeted treatments. And now, with tech that could pick up on patterns in our speech, eye movements and facial expressions, could machines be the future in global health?
If you’ve ever asked Siri a question, been motivated to run by a nagging watch, or suddenly realised you were flirting with a charming chatbot, you know our daily interactions, our access to information, our choices, are increasingly mediated by machines.
And, if you’re anything like me, (Co-Director of Common Thread, global health communications), it feels less like: “the Robots are coming!!!” and more like, “damn, when did they get here, and who invited them?”
Algorithms, pattern recognition, neural networks and deep learning — all part of the broader family of machine learning and artificial intelligence — will only continue to expand their influence on our lives. Our friends at ideas42 think “machine learning is poised to revolutionise the application of behavioural science.”
Whether or not this revolution is coming to save humanity or dismantle it, is a debate that feels a little on the lines of Will Smith’s ‘I, Robots’ film plot.
Trending AI Articles:
1. Making a Simple Neural Network
2. Neural networks for solving differential equations
Giving us a glimpse into the utopian and dystopian potential for AI in global health, Johanna Skilling of Burson Cohn & Wolfe, talked us through what happens when technology evolves to reinforce or replace human interactions in healthcare. Is empathy a purely human trait, or can it be engineered? And can engineered empathy produce better health outcomes than the human equivalent?
Common Thread: How could AI work in our field, in global health communications and how can we use it effectively?
Johanna: We’re all going to be affected by AI, robotics, pattern recognition and behaviour change mitigated by technology. Our responsibility is to make it serve us and our fellow human beings — for mental and physical care, for better lives.
WHO estimates a projected shortfall of 18 million health workers by 2030, mostly in low- and lower-middle income countries.
Shortages are also being compounded by the difficulties of deploying healthcare workers in rural, remote and under-served areas.
Common Thread: Could AI plug the massive global healthcare worker shortage?
Johanna: If you’re sick, if you’re in pain or if you need something, then who’s going to be available? If you’re lucky enough to see a nurse, a doctor or a technician, they’re going to be even more overwhelmed than they are now.
In the future, that doctor who now has had seven minutes with a patient might have five, maybe they’ll have three in a developing country, maybe in an emergency situation they’ll have 30 seconds, to ask ‘what do you have?’ or ‘how do I triage you’?
As a human being when you’re sick or displaced, or worried, empathy not only has an immediate mental health effect, but it also can have a difference in outcomes. If I’ve 30 seconds with you, how much do I know? And do I send you to the right place?
For women’s health which I’m more familiar with, it’s very easy for doctors, even female doctors when they’re rushed, to say, ‘it’s nothing’ or ‘it’s in your head’, or ‘you’ll be fine’ — and you won’t be fine.
Common Thread: So how can empathy be programmed into these outcomes?
Johanna: We give off biometric signals.
Researchers have isolated 27 different sets of expressions. So, say we’re concerned, worried, scared , or feeling joyous — these biometric signals can be processed and can be turned into pattern recognition.
It’s the same thing with tone of voice, breathing, body language, and things we’ve not yet considered.
There’s lots of pattern recognition available to ask ‘can machines emulate an empathetic human?’
“Machines will never be empathetic
but they can mimic empathy for
better patient experiences and outcomes.”
Johanna Skilling
Common Thread: So how transformative could the combination of data and AI be for global health?
Johanna: Pattern recognition could be used for behavior change.
Data and intelligence can pinpoint the need, bringing aid to potentially more people and places. Data can ask, ‘Is a particular disease status on the rise?’ ‘Are diagnoses going up?’ ‘Are healthcare workers going down?’
Common Thread: Aren’t facial expressions and body language so culturally specific? Could AI misdiagnose the signals from one culture to another?
Johanna: Pattern recognition comes from hard-wired facial tics. What the research has shown is that there are some cultural differences but there is a lot of commonality. So what you can do is say in Eastern cultures, certain facial patterns mean one thing and in Western cultures another and so on. So while there are cultural differences, there are underlying similarities.
These cultural differences can also be mapped.
Common Thread: So how could public health organisations apply this mapping in the future — say for instance in preventing the spread of a disease in a remote location, where there are very few healthcare workers?
Johanna: It’s completely possible to use AI or programme an avatar to be a certain type of person, of a particular age or gender, with a certain dialect.
For instance, if you’re in rural Afghanistan you could work with a developer and someone skilled in that dialect and pilot a public health initiative in that village in a matter of days, working with the population to improve it.
I think you could do wonders.
Common Thread: And so do we! We’re excited about how we can continue to use technology and AI as it develops in our work to change human behaviour to stop diseases from spreading.
If you’re interested in other insights from the fields of global health, behavioural science and designing for change, check out the latest edition of Common Thread’s Newsletter, The Stitch, or listen to the full interview with Johanna Skilling and other experts in their field on our SoundCloud Channel.
