December 22, 2025 • 36 mins
What if our smartest health tools still miss the people who need them most? We sit down with AI and digital health scientist Mireille Gomes to examine how innovation can serve dignity, not just efficiency—and what it takes to build technology that works from Geneva to rural clinics without electricity.
The journey of Mireille Gomes spans continents and roles, from vaccine strategy at Gavi to AI diagnostics at Merck. Together, we unpack the real barriers to deployment—uneven infrastructure, overworked staff, and data voids that erase entire communities from the record. We look at consent‑first design, why open data must be truly anonymous, and how representation in civil registration and vital statistics underpins every “fair” algorithm. You’ll hear pragmatic ideas for triage tools that flag urgency in seconds, health education in local languages, and micro‑local models that adapt to context while sharing standards globally.
We also push on the hard questions: Who decides which data matters? Can algorithms be biased toward justice if the world is not? Where is the line between breakthrough and overreach when crises demand speed? Mirielle argues for building abuse cases into development, testing for misuse before launch, and preserving community storytelling—especially Indigenous knowledge—alongside dashboards. The goal is health equity by design, so no one’s care depends on their birthplace or bandwidth.
If you care about AI in healthcare, data justice, and solutions that actually work on the ground, this conversation offers a clear roadmap and candid guardrails. If it resonates, subscribe, leave a review, and share it with someone shaping the future of digital health.
Check out "Protection for the Inventive Mind" – available now on Amazon in print and Kindle formats.
The views and opinions expressed (by the host and guest(s)) in this podcast are strictly their own and do not necessarily reflect the official policy or position of the entities with which they may be affiliated. This podcast should in no way be construed as promoting or criticizing any particular government policy, institutional position, private interest or commercial entity. Any content provided is for informational and educational purposes only.
Listen
Watch
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 4 (00:00):
The principle would be if you can identify both the
positives and the harmfulbehaviors a particular
technology can cause, that youdo everything possible to
prevent the harmful behaviorfrom happening.
Speaker 1 (00:24):
Can algorithms heal?
Reimagining health equity withAI and data justice.
So today we are joined bysomeone who proves that science
and strategy don't have to livein separate worlds.
Miral Gomes is an AI anddigital health scientist with a
(00:44):
PhD from Oxford, in my case, theother place, I'm from
Cambridge.
Global experience in healthsystems reforms and a deep
commitment to reducing healthinequities.
From leading vaccine strategyat GAVI to building AI
diagnostic at Merck, she hasworked across continents,
(01:07):
institutions, and sectors tomake data-driven innovation
serve the public.
Whether negotiating with globalalliances or coding genomics
models, Mural brings bothprecision and purpose to
everything she touches.
In this episode, we ask, can wedesign AI system that not only
(01:30):
predicts disease, but alsoprotect dignity?
What does justice look like ina world where health is
increasingly meditated by data?
So welcome to our podcast.
And please uh give us uh um ashort introduction of how you
landed here in Geneva.
Speaker 4 (01:51):
Thank you for having me and that very gracious
introduction.
Um I was working in differentparts of Africa and Asia, and I
wanted to understand how globalpolicy was made that was
influencing all the healthdecisions in the countries I was
(02:13):
working with.
And since a lot of the fundingcame from Geneva, I thought it
was time for me to come here andunderstand how these decisions
were made.
So I came to Geneva for a jobat GAVI, the Vaccine Alliance.
And I've been here for a while,both at GAVI and at Merck, and
it's interesting to see theworld from this lens.
Speaker 1 (02:36):
Yeah, it's Geneva is a beautiful ecosystem.
Yes.
And it's full of people fromeverywhere, from every
background.
So it's always a um I f I fellin love with the place before I
fell in love with my husband,who is a Genoa.
So yes, I completely see thethe beauty of it, and also
working in in such a beautifulfield, health, and also very um
(02:59):
taxing, very important, very uhvery impactful.
So um I'm very happy to haveyou here on the podcast and
let's dive in, please.
So you work in digital healthfrom Canada to Ethiopia, from
policy making to algorithmdesign.
What led you to combinescientific research with public
(03:20):
strategy?
Speaker 4 (03:21):
Um, so a bit of background about myself.
I was born in India.
My family then moved to Omanand then Canada.
I'm Canadian by citizenship,uh, all for better quality of
life.
And along the way, I've movedand lived in different
countries.
For me, doing this life journeyand seeing the health
(03:42):
inequities simply because um ofyour birth lottery and just how
where you're born determinesyour access to health and
dignity of life made me want tocombine my scientific training
with public strategy to enablehealth equity.
Speaker 1 (04:02):
Yeah, it makes uh um when when you have the
opportunity of um coming from adeveloping country, moving to a
developed country, you see somuch, your your mind's changing
so much.
And and you see that thingsthat for you may have been um
normal.
I come from the DominicanRepublic, so it's also a
developing country, that uh somesystematic failures are
(04:27):
accepted as the norm.
And then you you go abroad andyou realize, well, this isn't
this is not the norm, and thisis not the norm here, and it
shouldn't be the norm there.
So you start questioning a lotthe system and what can you do
to improve it.
Speaker 4 (04:42):
And I I completely agree, and there's also a lot of
history and of colonization andother forces that led to the
current situation in developingcountries.
So I personally do not feellike it is a self-created
problem, but rather a globalproblem for global citizens,
(05:04):
since everyone has been impactedby people across other borders
throughout their history, andit's not their own fault or
responsibility of why things arethe way they are.
So um, as global citizens, weare responsible for each other.
Speaker 1 (05:22):
Yeah, and that's why we we built these amazing
organizations to find waystogether to tackle all these
issues that affect us all.
Yes, agreed.
So you have led large-scalepartnerships at Merck and Gabi.
What's the biggest challengewhen using AI or digital health
tools across very differenthealth systems?
Speaker 4 (05:45):
For me, uh there are two specific challenges.
One is um the training ofhealth workers and two,
infrastructure to give someexamples.
Um, with infrastructure, youcould build an AI tool for a
developed country hospital thatallows a highly trained surgeon
(06:09):
to make a very precise incisionduring surgery, and the AI could
be great in guiding the surgeonto be minimally invasive.
In contrast, there are lots ofhospitals that don't even have
electricity.
So, how do you build an AI toolwith non-functioning
(06:31):
electricity or smartphones oranything else?
So for me, that's one of thehuge discrepancies in preventing
AI from being accessible in allsettings.
Um, and the second one is thetraining of staff.
So going back to my surgeon,you could have an Oxford or
(06:53):
Cambridge trained surgeon.
Um, or you could have thecommunity health worker who has
just literacy skills, who's beentaught the basic nursing, et
cetera, training, and thenyou're dumping them with a
complicated digital tool on topof their already heavy task
load.
(07:14):
Um, so these are some of theconsiderations when developing
digital tools that have to bekept in mind, both in terms of
will it actually function in thesetting if there's no
electricity?
And two, like you're askingsomebody with basic reading and
maybe basic math skills to lookat your very complicated tool to
(07:38):
use it.
Speaker 1 (07:39):
So it's is the environment side and also the
person behind the technologythat you are um asking them to
learn or to use and or totackle.
Exactly.
So it's it's it's to keep inmind the realities that are
different.
Yes.
And every country has their owncontext and their own um
strength and also weaknesses.
Speaker 4 (08:00):
And even within each country, like I could be at a
very expensive private hospitalin India, or I could be in an
area that has a mobile truckcoming to provide clinical
services.
Speaker 1 (08:12):
Yeah, it can it can differ so much even within the
country, yeah.
Yeah, yeah, makes makes perfectsense.
But who decides what datamatters or how it is used?
Speaker 4 (08:42):
That's a very good question.
Um I think the current data gapis that data is not
representative and doesn'tcapture all the uh groups of
individuals that should berepresented.
And I think that leads tobiases in the algorithm for
(09:03):
over-predicting orunderpredicting a certain um uh
person or type of character.
Um, so having firstly makingsure that every human on this
planet is counted.
That's just a basic human rightand dignity, and then having
(09:26):
that reflected in the data beingused by AI and machine learning
would be the holistic approach.
Um, there are a lot of placeswhere people don't have birth
certificates or deathcertificates.
Speaker 1 (09:42):
So they don't exist.
Speaker 4 (09:43):
They don't exist.
In paper.
In paper.
And it's very sad to not tohave walked this earth and have
not been recorded as existing.
Speaker 1 (09:56):
There's no footprint of what you have to.
Yeah.
Well, yeah, that's that'sthat's quite uh powerful to be
invisible in the system becausethen the system doesn't know
that you're there, so it doesn'tit doesn't take you take your
realities or your necessitiesinto account.
Speaker 4 (10:12):
And then how would an AI take it into account?
Speaker 1 (10:15):
The AI cannot go to the field and gather
information.
Speaker 4 (10:18):
Yes, and if the information doesn't exist, then
they you're underrepresented forthe AI too.
Speaker 1 (10:24):
Oh wow, that's a that's a scary thought.
Speaker 4 (10:27):
Yes, yes, unfortunately it's a lot of
people's reality.
Speaker 1 (10:30):
Yeah, yeah, makes sense.
Um, you work on diseaseprediction, vaccine strategy,
and data access policy.
Where do you draw the linebetween innovation and
overreach?
Speaker 4 (10:44):
So I think we are an interesting species as humans.
Speaker 1 (10:49):
That's a beautiful way of putting it.
Speaker 4 (10:51):
I think we're driven by both curiosity and progress.
And um, so I think as a speciesthere will always be both
overreach and innovation, that'sjust because who we are.
But um, as the amazingSpider-Man comics say, with
great power comes greatresponsibility.
(11:14):
Um, and that was originallysaid by Voltaire.
I think Spider-Man took it fromthere.
But um I think yeah, there'salways going to be both
innovation and overreach, andyou just have to be responsible.
Speaker 1 (11:26):
And find a balance.
Yes.
Which is not very easy.
It's very it's is it's easy tosay in the abstract, but in the
reality is it's it's hard tofind a balance and also to take
into account the differentforces as well that drive the
innovation or can or it canhinder innovation as well.
Speaker 4 (11:46):
Agreed.
And I I think with anycommodity or power, it's based
on the individual and how theyuse it.
Speaker 1 (11:54):
Yeah.
Yeah, it's a tool.
It's a tool.
You can use it to build ordestroy it.
Destroy.
So you work at the highestlevels of scientific and
strategic design, but how do wemake sure the person on the
ground, a patient, a nurse, acommunity, is not left behind?
Speaker 4 (12:16):
Um so it's tough, especially because we just
talked about how not everybody'seven counted or documented.
But making sure as many of thedifferent groups of people are
represented in the design phaseand the design thinking before a
(12:37):
solution is scaled is importantto ensure the local context and
the gaps that maybe as adeveloped country person I just
enforce without even realizingwhat the problems are on the
ground.
So making sure that every groupis represented in the design
(12:59):
thinking would be a way to tryto capture.
I'm sure we will not beperfect, but at least it's an
attempt.
Speaker 1 (13:06):
Well, perfection doesn't exist.
That's the first thing.
Uh, but it's it makes sense tofrom the design, uh the the
first step is to understand whoare you gonna impact, how you're
gonna impact their lives andthe the communities and and and
the people in the in the healthof professionals.
So it's important for and fromthat point, then you can you can
(13:30):
design better, you can you canaddress the issues better
because you already have thefirst information, which I I I
believe that a lot of programsand projects they they should
take their time to to go to thereality, understand it, and
design from that reality.
Speaker 4 (13:47):
And I think humility is also important.
Yes.
Like not designing by thinkingyou have all the answers.
Exactly.
I think the innovation is oftenwith the people on the ground
and we fail to appreciate it.
Speaker 1 (14:01):
And they already know what is needed as well.
They have the a deepunderstanding of what's the
issue, what what what exactlyneeds to be changed, and they
maybe have an idea how to changeit, but they don't have the
resources.
Speaker 4 (14:13):
Completely agree.
And uh maybe I'm just helpingthem get the resources to
implement their innovation.
Speaker 1 (14:22):
Pushing very bit.
So what does health equity bydesign mean to you?
Speaker 4 (14:31):
Um I think designing health systems to ensure access
to health care and dignity foreveryone is health equity to me.
Um, there shouldn't be anybodyon this planet who shouldn't
have full access to healthcaresimply because they were born in
(14:51):
the wrong house or in the wrongdistrict.
And it's sad to think that weas humans would allow another
human to live an undignifiedlife.
Speaker 1 (15:06):
Yeah, just just because they happen to be in the
wrong side of the tracks.
Speaker 4 (15:10):
Yes, exactly.
And um yeah, I think if youhave the ability, the resources,
then it's also your dignity andresponsibility to help those
who don't.
Speaker 1 (15:29):
So and synchronous change.
Yeah.
You may be privileged in oneone moment and then the next you
may may not, because lifechanges and takes you to a
different path.
Speaker 4 (15:39):
Agreed.
Speaker 1 (15:40):
Yeah.
So it's important to see that.
Can algorithms be biasedtowards justice, or do we need
more than just better code?
Speaker 4 (15:51):
So a lot of algorithms are based on the data
they're provided, and we talkedabout having better
representation in the data toensure that everyone is counted.
Um and yes, you could buildinto algorithms some principles
of justice and equity, but aslong as there are data gaps, it
(16:14):
will be difficult to ensureequity.
Speaker 1 (16:17):
Yeah.
Because it's this is it comesfrom the core of the
information.
Yes.
Yeah.
Yeah.
If you could redesign one partof the global health
infrastructure with AI, wherewould you start and what will be
different?
Speaker 4 (16:32):
So this is a difficult question.
Speaker 1 (16:35):
I would just solve the world problems.
Speaker 4 (16:40):
Um I think, especially given the current
climate of lack of global healthfunding or stopping global
health funding by decisions byvery few people.
Um I think AI could help in acontext where you have to do
(17:01):
more with less.
So now where we're fundingconstrained, etc., if there are
already, even before thisfunding crisis, there were
already a lot of hospitals orprimary care that didn't have
sufficient staff.
Um, if there's a way that AIcan help triage patients to just
(17:25):
basic uh kind of primary,secondary, tertiary care, if um
AI could bring, for example,MPOX epidemic is still going on,
it's not in our news anymore,but just health education in
local languages or promotion ofgood practices, um, just simple
(17:49):
ways to fill the gaps that don'trequire a very complicated
algorithm, but just umidentifying the gaps and
providing the information.
Speaker 1 (18:03):
Provides efficiency, incentive incentive, I guess
they're critical and takes time.
Yes.
Because when you receive apatient, the first thing you
need to assess is how um how canI help the patient?
Yeah, and how urgent is theneed that they uh that they
have.
So and that takes time toevaluate.
Speaker 4 (18:20):
And I think that's true across every health system.
Um, so something that couldwork for the NHS in the UK,
which is extremely constrainedat the moment, to help identify
the patient's um severity ofdisease and where they should
go, could also help in a localvillage in Sierra Leone.
Yeah, of course.
(18:41):
Uh, where the health workerdoesn't only have a high school
education, but the AI could atleast identify the symptoms and
say this is high priority or youneed to be treated now.
Speaker 1 (18:54):
Yeah, exactly.
Speaker 4 (18:55):
Exactly.
unknown (19:04):
Okay.
Speaker 1 (19:06):
So now we're going into a bit of a fun part of the
session after all this heavytalk.
I'm ready.
Uh flash uh questions.
Okay.
So you have to pick one.
Um with pick one withoutnecessarily uh saying why, but
if you want to explain yourselfa little bit, it's okay.
(19:26):
We're open to conversation.
unknown (19:29):
Got it.
Speaker 1 (19:30):
But the idea is to uh think quick.
Speaker 4 (19:34):
I'm ready.
Speaker 1 (19:34):
Okay.
Diagnosing early orunderstanding deeply.
Diagnosing early.
Open source health data ortightly protected patient
sovereignty.
Speaker 4 (19:50):
Open source health data with full anonymity so that
it cannot be traced to anindividual.
Speaker 1 (19:58):
Perfect.
Tech that scales or localizetools that adapt.
Speaker 4 (20:35):
Tech uh that scales with the ability to customize
locally.
High rate.
Yes.
Speaker 1 (20:44):
I was like, can I get away with that answer?
Data transparency at all costsor consent uh consent first
design, even if it slows down uhthe system.
Speaker 4 (20:59):
Consent first.
Speaker 1 (21:02):
A world where everyone is tracked for health
or where they are leftunmonitored.
Um opt-in.
Okay.
Global health driven by AIdashboards or shaped by
community storytelling.
Speaker 4 (21:21):
Global health AI dashboards with community health
storytelling.
Speaker 1 (21:26):
Another hybrid.
Another hybrid.
Can I elaborate?
Yeah, yeah, of course.
Speaker 4 (21:30):
Um so during the pandemic, a lot of Aboriginal
communities in Canada andelsewhere lost a lot of their
elders due to COVID.
And they're in that communityis the history of storytelling
plays a major role.
And it's a passed down of notonly um traditional healing
methods, but also of history andancestors and all.
(21:54):
So losing that completely is aloss to many generations and
several.
So having captured thesestories so that they can be
passed down digitally orotherwise is important.
Speaker 1 (22:09):
Yeah.
So it's it it it has a a wealthof uh of information that we're
losing.
Yeah, that we're losing everytime that uh because uh every
time we lose uh a communityleader, elder that have seen so
much and knows so much and alsohas has learned from his own
elders.
(22:29):
So it's this this kind ofinformation, especially in
communities, indigenous uhpeople's communities, where um
they they have such uh uh a deepunderstanding far beyond that
we can we can grasp of of life,of health, of of um everyday um
information as well,environment, um, agriculture is
(22:52):
is it's limitless theinformation that they have.
So it's it's uh it's importantnot to lose it and not to let it
just slip away.
Speaker 4 (23:00):
And if the digital communications or others can
help capture that, that would bea way to preserve.
Speaker 1 (23:08):
Yeah.
Always with their consent, ofcourse.
Yes, of course.
Then uh next question (23:11):
a universal health algorithm or
millions of micro-local ones.
Speaker 4 (23:21):
Micro-local ones for the context.
Speaker 1 (23:25):
Faster breakthroughs or longer-term resilience?
Speaker 4 (23:31):
Um risk versus benefit.
Okay.
So, for example, during thepandemic, we had to, I was at
Gabby, that was uh you had toget vaccines out fast.
Um, and then you weigh the riskof getting a vaccine out as
soon as possible with versus theshort-term and long-term side
effects.
Um, but these areconsiderations that are taken
(23:55):
into account during the decisionmaking, and regulatory bodies
also help ensure that risks areminimized.
Speaker 1 (24:03):
And I think in that in that aspect, there's no um
easy answer.
There's no easy answer becausethere's there's always gonna be
a lost.
Yes.
And you need to wait which whatwhich one are you willing to to
go ahead and and live with?
Yeah.
So it's no, it's an impossiblesituation, it's an impossible
(24:24):
situation.
Yeah, very tough question.
Yeah, yeah, yeah.
Okay, last flash question.
Could AI train on clinicaltrials or on lived experiences?
Speaker 4 (24:39):
Um both.
I mean the clinical trials.
Yes.
There's no way you can get allof the information during a
clinical trial.
Um the real world evidence oncea drug is on the market in a
not so controlled setting isalso important.
(25:00):
So mm losing any of thatinformation would be tragic.
Speaker 1 (25:05):
Yeah, yeah, it's it's because then when one is
outside, you have the wholeworld that can tell you how it
works.
Speaker 4 (25:13):
And maybe you haven't considered a specific patient
group in your clinical trial.
For example, a lot of pregnantmothers, etc., not to say that
the drug is given to them after,but there are maybe different
minority populations that youhaven't included, whether
racially or disease-specificminorities.
(25:33):
So tracking the information onhow it's affecting these
different groups is veryimportant because we're seeing
in healthcare now, uh afterhaving years and years of data,
we are seeing that drugs mightaffect different minority groups
differently.
So making sure that you try tocapture them in your clinical
(25:57):
trials to begin with, but thenalso after.
Speaker 1 (26:01):
Yeah, there's so many factors.
There's um um the medicinecombination as well.
So a person who has uh anotherunderlying condition that is
taking medicine for somethingelse that interacts with this
new medicine, there's also canbe a reaction for that, also a
genetical predisposition, umeven the the the the health uh
(26:24):
um um uh habits of the person.
How likely is the person to eatwell, how the the food
combination, there even kind offood that can also have a bad
reaction to medicine.
Speaker 3 (26:36):
Yes.
Speaker 1 (26:36):
So it's there's so many factors to take into
account that yeah, not nothingtrumps uh uh real life.
Speaker 4 (26:43):
Agreed, and we recently saw that from the
obesity drugs that have beenidentified where they were
originally not designed forobesity but other diseases, and
then the uh post-marketinformation has shown that they
affect obesity as well, and nowthey've become blockbuster
(27:06):
drugs.
Uh there are a few I shouldn'tshow any biases.
Speaker 1 (27:12):
Yeah, I'm just asking for a friend.
unknown (27:15):
Of course.
Speaker 1 (27:18):
Uh yes, this that's uh and also the way that
interacts with you, uh it can bealso so so different.
Yes.
Um, I have a very low thresholdin any medicine in general
because I I I hate takingmedicine.
So if I can avoid it, I willendure a headache.
Oh wow, stomachache.
I I I prefer to I make a tea, Idrink some water.
(27:39):
Okay.
I I try to not to take a lot ofmedicine, but when I do, I'm
I'm I'm down.
It really is.
It's you're at your last.
Like it's like it's it's umit's crazy because I think it's
because my body's not used tohaving them.
So once in mysticism, it'slike, okay, I'm gonna grab every
single one drop of molecule,and you're gonna get it.
(28:03):
Uh and something funny happenedto to light up the mood.
Um, the first day that I hadwith my husband, uh, we were
supposed to meet uh for for tea.
And I fell asleep because Itook uh um um a pill for my cold
that I had.
I was I had a very, very badcold, and I never take anything.
(28:25):
So I decided to take it, and Ifell completely asleep.
Completely, and because he Ididn't know, but he had a little
bit of anti-allergic into it.
And I also was down.
The poor guy waited for me forover an hour.
Speaker 4 (28:40):
Oh wow, but we met.
Speaker 1 (28:42):
He's a keeper, and he's Swiss, so yeah, being on
top is very important.
So yeah, he really uh uh it'sit's something funny that I say
uh all the time.
Like, I am I'm I'm very much uhI could be a very interesting
specimen.
You would be in this minoritygroup of people.
(29:04):
You would be like, what iswrong with her body that takes
on so many medicines?
Speaker 4 (29:11):
See, this is why we should track post market falling
asleep for an hour.
Speaker 1 (29:15):
There you go, for just a common cold medicine.
Okay, now uh the next the nextgame.
You take the pile there, andthe question um I will ask you
the statement.
You will tell me if it's true,meaning that you believe it's
happening right now and it'sabout to happen, okay, or
(29:37):
futuristic with a tiny robot.
That means that it's somethingsci-fi is gonna happen way in
the future, or it may neverhappen.
Speaker 4 (29:46):
Okay.
Speaker 1 (29:47):
Okay, ready?
Ready.
Health systems will use AI todetect disease outbreaks before
symptoms start to appear.
Patients will have digitalhealth twins that simulate their
(30:08):
care outcomes.
Groups of patients.
Okay, not individual groups,yes.
Okay.
You'll be able to pause yourdigital health footprint.
Speaker 4 (30:23):
Just no.
Neither true or futuristic.
Just no.
No, no pausing.
Speaker 1 (30:29):
No pausing.
Public health policies will beco-drafted by humans and AI.
Governments will use AI topredict well healthcare workers
should be sent before crisishappened.
Diagnosis will include anemotional impact score.
(30:53):
Possible.
Yeah, but it would be nice.
You have the right to deleteyour medical prediction history.
Speaker 4 (31:08):
It it depends on the legislation where you are.
Speaker 1 (31:12):
And what what's what the service uh the purpose is
gonna serve.
Yes.
Because they need informationto treat you.
Speaker 4 (31:18):
And if you don't have it, then yes, but it's still
your information, so you shouldhave a right to it.
Speaker 1 (31:25):
To be forgotten.
Speaker 4 (31:26):
Yeah.
Speaker 1 (31:27):
Yeah.
Tricky, tricky.
Yes, I'll just leave it here.
No question, no answer.
No answer.
AI will be able to detect biasin global health funding models.
unknown (31:39):
True.
Speaker 1 (31:39):
True.
Health insurance will bepartially determined by your
microbiote.
Speaker 4 (31:46):
Depends on legislation.
Speaker 1 (31:49):
So no answer.
Yes.
And lastly, every hospital AIwill have a built-in system that
checks if its decisions arefair and ethical before taking
action.
Speaker 4 (32:03):
So I mean, even between you and I, we might have
a different interpretation ofwhat's fair and ethical.
Speaker 1 (32:09):
Yes.
Speaker 4 (32:09):
So it would be hard to code that into an algorithm.
Speaker 1 (32:13):
So no answer.
No answer.
Love that, love that.
Yeah, neutrality.
Yes.
Speaker 4 (32:19):
And Switzerland.
Speaker 1 (32:20):
Yeah, yeah, we are in Switzerland neutral.
So, final question.
If you could embed oneinvisible value into every
health innovation being builtright now, not a feature, not a
gadget, but a principle.
What would it be and why?
Speaker 4 (32:41):
Um the principle would be if you can identify
both the positives and theharmful behaviors a particular
technology can cause, that youdo everything possible to
prevent the harmful behaviorfrom happening.
Um, so just some examples.
(33:03):
Historically, um during WorldWar II, the first early
computers were used to takecensus information to quickly
identify ethnic minority groupsand round everyone up.
Um with latest AI technologyfor images, it's being used for
(33:26):
lots of interesting things, butit's also being used for child
pornography.
Um so if these harmful outcomescan be identified by the tech
developers before the technologyis released, then putting
precautionary measures in thetechnology, I think, would be a
(33:49):
principle that um would be aresponsibility of the developer.
Um on the other hand, it's notall doom and gloom.
There's also lots of beautifulthings that technology is being
used for, for example, to savethe Amazon forest and identify
deforestation or rehouse 60,000families after the Nepalese
(34:13):
earthquake.
So there's a lot of beautifulthings tech is being used for,
but for the harmful things, justlike Einstein's relativity
theory of the atom that led tothe atomic bomb versus
everything else, I think theyshould be identified, and every
measure that can be taken toprevent the harmful behavior
from happening is theresponsibility of the developer.
Speaker 1 (34:36):
Perfect.
So to put the fences in theright places.
Yes, where possible.
Thank you so much.
Thank you for your time.
Thank you for coming andtalking uh with us.
It's been uh quite a pleasureuh to learn uh about the work
that you do and also uheverything that is um uh the
(34:58):
potential that technology canhave in health.
And thank you for reminding usthat digital health is not about
faster systems or smartertools, it's about people, it's
about justice, it's about askingbetter questions before we
scale solutions.
Because health is not just adata point, it's a human rate.
Speaker 4 (35:21):
Thank you for having me.
It's been a pleasure.
Speaker (35:28):
Thank you for listening to Intangibia, the podcast of
Intangible Law.
Plain talk about intellectualproperty.
Did you like what we talkedtoday?
Please share with your network.
Do you want to learn more aboutintellectual property?
Subscribe now on your favoritepodcast player.
Follow us on Instagram,Facebook, LinkedIn, and Twitter.
(35:49):
Visit our websitewww.intangibilia.com.
Copyright Letista Caminero2020.
All rights reserved.
This podcast is provided forinformation purposes only.
The principle would be if you can identify both the
positives and the harmfulbehaviors a particular
technology can cause, that youdo everything possible to
prevent the harmful behaviorfrom happening.
Speaker 1 (00:24):
Can algorithms heal?
Reimagining health equity withAI and data justice.
So today we are joined bysomeone who proves that science
and strategy don't have to livein separate worlds.
Miral Gomes is an AI anddigital health scientist with a
(00:44):
PhD from Oxford, in my case, theother place, I'm from
Cambridge.
Global experience in healthsystems reforms and a deep
commitment to reducing healthinequities.
From leading vaccine strategyat GAVI to building AI
diagnostic at Merck, she hasworked across continents,
(01:07):
institutions, and sectors tomake data-driven innovation
serve the public.
Whether negotiating with globalalliances or coding genomics
models, Mural brings bothprecision and purpose to
everything she touches.
In this episode, we ask, can wedesign AI system that not only
(01:30):
predicts disease, but alsoprotect dignity?
What does justice look like ina world where health is
increasingly meditated by data?
So welcome to our podcast.
And please uh give us uh um ashort introduction of how you
landed here in Geneva.
Speaker 4 (01:51):
Thank you for having me and that very gracious
introduction.
Um I was working in differentparts of Africa and Asia, and I
wanted to understand how globalpolicy was made that was
influencing all the healthdecisions in the countries I was
(02:13):
working with.
And since a lot of the fundingcame from Geneva, I thought it
was time for me to come here andunderstand how these decisions
were made.
So I came to Geneva for a jobat GAVI, the Vaccine Alliance.
And I've been here for a while,both at GAVI and at Merck, and
it's interesting to see theworld from this lens.
Speaker 1 (02:36):
Yeah, it's Geneva is a beautiful ecosystem.
Yes.
And it's full of people fromeverywhere, from every
background.
So it's always a um I f I fellin love with the place before I
fell in love with my husband,who is a Genoa.
So yes, I completely see thethe beauty of it, and also
working in in such a beautifulfield, health, and also very um
(02:59):
taxing, very important, very uhvery impactful.
So um I'm very happy to haveyou here on the podcast and
let's dive in, please.
So you work in digital healthfrom Canada to Ethiopia, from
policy making to algorithmdesign.
What led you to combinescientific research with public
(03:20):
strategy?
Speaker 4 (03:21):
Um, so a bit of background about myself.
I was born in India.
My family then moved to Omanand then Canada.
I'm Canadian by citizenship,uh, all for better quality of
life.
And along the way, I've movedand lived in different
countries.
For me, doing this life journeyand seeing the health
(03:42):
inequities simply because um ofyour birth lottery and just how
where you're born determinesyour access to health and
dignity of life made me want tocombine my scientific training
with public strategy to enablehealth equity.
Speaker 1 (04:02):
Yeah, it makes uh um when when you have the
opportunity of um coming from adeveloping country, moving to a
developed country, you see somuch, your your mind's changing
so much.
And and you see that thingsthat for you may have been um
normal.
I come from the DominicanRepublic, so it's also a
developing country, that uh somesystematic failures are
(04:27):
accepted as the norm.
And then you you go abroad andyou realize, well, this isn't
this is not the norm, and thisis not the norm here, and it
shouldn't be the norm there.
So you start questioning a lotthe system and what can you do
to improve it.
Speaker 4 (04:42):
And I I completely agree, and there's also a lot of
history and of colonization andother forces that led to the
current situation in developingcountries.
So I personally do not feellike it is a self-created
problem, but rather a globalproblem for global citizens,
(05:04):
since everyone has been impactedby people across other borders
throughout their history, andit's not their own fault or
responsibility of why things arethe way they are.
So um, as global citizens, weare responsible for each other.
Speaker 1 (05:22):
Yeah, and that's why we we built these amazing
organizations to find waystogether to tackle all these
issues that affect us all.
Yes, agreed.
So you have led large-scalepartnerships at Merck and Gabi.
What's the biggest challengewhen using AI or digital health
tools across very differenthealth systems?
Speaker 4 (05:45):
For me, uh there are two specific challenges.
One is um the training ofhealth workers and two,
infrastructure to give someexamples.
Um, with infrastructure, youcould build an AI tool for a
developed country hospital thatallows a highly trained surgeon
(06:09):
to make a very precise incisionduring surgery, and the AI could
be great in guiding the surgeonto be minimally invasive.
In contrast, there are lots ofhospitals that don't even have
electricity.
So, how do you build an AI toolwith non-functioning
(06:31):
electricity or smartphones oranything else?
So for me, that's one of thehuge discrepancies in preventing
AI from being accessible in allsettings.
Um, and the second one is thetraining of staff.
So going back to my surgeon,you could have an Oxford or
(06:53):
Cambridge trained surgeon.
Um, or you could have thecommunity health worker who has
just literacy skills, who's beentaught the basic nursing, et
cetera, training, and thenyou're dumping them with a
complicated digital tool on topof their already heavy task
load.
(07:14):
Um, so these are some of theconsiderations when developing
digital tools that have to bekept in mind, both in terms of
will it actually function in thesetting if there's no
electricity?
And two, like you're askingsomebody with basic reading and
maybe basic math skills to lookat your very complicated tool to
(07:38):
use it.
Speaker 1 (07:39):
So it's is the environment side and also the
person behind the technologythat you are um asking them to
learn or to use and or totackle.
Exactly.
So it's it's it's to keep inmind the realities that are
different.
Yes.
And every country has their owncontext and their own um
strength and also weaknesses.
Speaker 4 (08:00):
And even within each country, like I could be at a
very expensive private hospitalin India, or I could be in an
area that has a mobile truckcoming to provide clinical
services.
Speaker 1 (08:12):
Yeah, it can it can differ so much even within the
country, yeah.
Yeah, yeah, makes makes perfectsense.
But who decides what datamatters or how it is used?
Speaker 4 (08:42):
That's a very good question.
Um I think the current data gapis that data is not
representative and doesn'tcapture all the uh groups of
individuals that should berepresented.
And I think that leads tobiases in the algorithm for
(09:03):
over-predicting orunderpredicting a certain um uh
person or type of character.
Um, so having firstly makingsure that every human on this
planet is counted.
That's just a basic human rightand dignity, and then having
(09:26):
that reflected in the data beingused by AI and machine learning
would be the holistic approach.
Um, there are a lot of placeswhere people don't have birth
certificates or deathcertificates.
Speaker 1 (09:42):
So they don't exist.
Speaker 4 (09:43):
They don't exist.
In paper.
In paper.
And it's very sad to not tohave walked this earth and have
not been recorded as existing.
Speaker 1 (09:56):
There's no footprint of what you have to.
Yeah.
Well, yeah, that's that'sthat's quite uh powerful to be
invisible in the system becausethen the system doesn't know
that you're there, so it doesn'tit doesn't take you take your
realities or your necessitiesinto account.
Speaker 4 (10:12):
And then how would an AI take it into account?
Speaker 1 (10:15):
The AI cannot go to the field and gather
information.
Speaker 4 (10:18):
Yes, and if the information doesn't exist, then
they you're underrepresented forthe AI too.
Speaker 1 (10:24):
Oh wow, that's a that's a scary thought.
Speaker 4 (10:27):
Yes, yes, unfortunately it's a lot of
people's reality.
Speaker 1 (10:30):
Yeah, yeah, makes sense.
Um, you work on diseaseprediction, vaccine strategy,
and data access policy.
Where do you draw the linebetween innovation and
overreach?
Speaker 4 (10:44):
So I think we are an interesting species as humans.
Speaker 1 (10:49):
That's a beautiful way of putting it.
Speaker 4 (10:51):
I think we're driven by both curiosity and progress.
And um, so I think as a speciesthere will always be both
overreach and innovation, that'sjust because who we are.
But um, as the amazingSpider-Man comics say, with
great power comes greatresponsibility.
(11:14):
Um, and that was originallysaid by Voltaire.
I think Spider-Man took it fromthere.
But um I think yeah, there'salways going to be both
innovation and overreach, andyou just have to be responsible.
Speaker 1 (11:26):
And find a balance.
Yes.
Which is not very easy.
It's very it's is it's easy tosay in the abstract, but in the
reality is it's it's hard tofind a balance and also to take
into account the differentforces as well that drive the
innovation or can or it canhinder innovation as well.
Speaker 4 (11:46):
Agreed.
And I I think with anycommodity or power, it's based
on the individual and how theyuse it.
Speaker 1 (11:54):
Yeah.
Yeah, it's a tool.
It's a tool.
You can use it to build ordestroy it.
Destroy.
So you work at the highestlevels of scientific and
strategic design, but how do wemake sure the person on the
ground, a patient, a nurse, acommunity, is not left behind?
Speaker 4 (12:16):
Um so it's tough, especially because we just
talked about how not everybody'seven counted or documented.
But making sure as many of thedifferent groups of people are
represented in the design phaseand the design thinking before a
(12:37):
solution is scaled is importantto ensure the local context and
the gaps that maybe as adeveloped country person I just
enforce without even realizingwhat the problems are on the
ground.
So making sure that every groupis represented in the design
(12:59):
thinking would be a way to tryto capture.
I'm sure we will not beperfect, but at least it's an
attempt.
Speaker 1 (13:06):
Well, perfection doesn't exist.
That's the first thing.
Uh, but it's it makes sense tofrom the design, uh the the
first step is to understand whoare you gonna impact, how you're
gonna impact their lives andthe the communities and and and
the people in the in the healthof professionals.
So it's important for and fromthat point, then you can you can
(13:30):
design better, you can you canaddress the issues better
because you already have thefirst information, which I I I
believe that a lot of programsand projects they they should
take their time to to go to thereality, understand it, and
design from that reality.
Speaker 4 (13:47):
And I think humility is also important.
Yes.
Like not designing by thinkingyou have all the answers.
Exactly.
I think the innovation is oftenwith the people on the ground
and we fail to appreciate it.
Speaker 1 (14:01):
And they already know what is needed as well.
They have the a deepunderstanding of what's the
issue, what what what exactlyneeds to be changed, and they
maybe have an idea how to changeit, but they don't have the
resources.
Speaker 4 (14:13):
Completely agree.
And uh maybe I'm just helpingthem get the resources to
implement their innovation.
Speaker 1 (14:22):
Pushing very bit.
So what does health equity bydesign mean to you?
Speaker 4 (14:31):
Um I think designing health systems to ensure access
to health care and dignity foreveryone is health equity to me.
Um, there shouldn't be anybodyon this planet who shouldn't
have full access to healthcaresimply because they were born in
(14:51):
the wrong house or in the wrongdistrict.
And it's sad to think that weas humans would allow another
human to live an undignifiedlife.
Speaker 1 (15:06):
Yeah, just just because they happen to be in the
wrong side of the tracks.
Speaker 4 (15:10):
Yes, exactly.
And um yeah, I think if youhave the ability, the resources,
then it's also your dignity andresponsibility to help those
who don't.
Speaker 1 (15:29):
So and synchronous change.
Yeah.
You may be privileged in oneone moment and then the next you
may may not, because lifechanges and takes you to a
different path.
Speaker 4 (15:39):
Agreed.
Speaker 1 (15:40):
Yeah.
So it's important to see that.
Can algorithms be biasedtowards justice, or do we need
more than just better code?
Speaker 4 (15:51):
So a lot of algorithms are based on the data
they're provided, and we talkedabout having better
representation in the data toensure that everyone is counted.
Um and yes, you could buildinto algorithms some principles
of justice and equity, but aslong as there are data gaps, it
(16:14):
will be difficult to ensureequity.
Speaker 1 (16:17):
Yeah.
Because it's this is it comesfrom the core of the
information.
Yes.
Yeah.
Yeah.
If you could redesign one partof the global health
infrastructure with AI, wherewould you start and what will be
different?
Speaker 4 (16:32):
So this is a difficult question.
Speaker 1 (16:35):
I would just solve the world problems.
Speaker 4 (16:40):
Um I think, especially given the current
climate of lack of global healthfunding or stopping global
health funding by decisions byvery few people.
Um I think AI could help in acontext where you have to do
(17:01):
more with less.
So now where we're fundingconstrained, etc., if there are
already, even before thisfunding crisis, there were
already a lot of hospitals orprimary care that didn't have
sufficient staff.
Um, if there's a way that AIcan help triage patients to just
(17:25):
basic uh kind of primary,secondary, tertiary care, if um
AI could bring, for example,MPOX epidemic is still going on,
it's not in our news anymore,but just health education in
local languages or promotion ofgood practices, um, just simple
(17:49):
ways to fill the gaps that don'trequire a very complicated
algorithm, but just umidentifying the gaps and
providing the information.
Speaker 1 (18:03):
Provides efficiency, incentive incentive, I guess
they're critical and takes time.
Yes.
Because when you receive apatient, the first thing you
need to assess is how um how canI help the patient?
Yeah, and how urgent is theneed that they uh that they
have.
So and that takes time toevaluate.
Speaker 4 (18:20):
And I think that's true across every health system.
Um, so something that couldwork for the NHS in the UK,
which is extremely constrainedat the moment, to help identify
the patient's um severity ofdisease and where they should
go, could also help in a localvillage in Sierra Leone.
Yeah, of course.
(18:41):
Uh, where the health workerdoesn't only have a high school
education, but the AI could atleast identify the symptoms and
say this is high priority or youneed to be treated now.
Speaker 1 (18:54):
Yeah, exactly.
Speaker 4 (18:55):
Exactly.
unknown (19:04):
Okay.
Speaker 1 (19:06):
So now we're going into a bit of a fun part of the
session after all this heavytalk.
I'm ready.
Uh flash uh questions.
Okay.
So you have to pick one.
Um with pick one withoutnecessarily uh saying why, but
if you want to explain yourselfa little bit, it's okay.
(19:26):
We're open to conversation.
unknown (19:29):
Got it.
Speaker 1 (19:30):
But the idea is to uh think quick.
Speaker 4 (19:34):
I'm ready.
Speaker 1 (19:34):
Okay.
Diagnosing early orunderstanding deeply.
Diagnosing early.
Open source health data ortightly protected patient
sovereignty.
Speaker 4 (19:50):
Open source health data with full anonymity so that
it cannot be traced to anindividual.
Speaker 1 (19:58):
Perfect.
Tech that scales or localizetools that adapt.
Speaker 4 (20:35):
Tech uh that scales with the ability to customize
locally.
High rate.
Yes.
Speaker 1 (20:44):
I was like, can I get away with that answer?
Data transparency at all costsor consent uh consent first
design, even if it slows down uhthe system.
Speaker 4 (20:59):
Consent first.
Speaker 1 (21:02):
A world where everyone is tracked for health
or where they are leftunmonitored.
Um opt-in.
Okay.
Global health driven by AIdashboards or shaped by
community storytelling.
Speaker 4 (21:21):
Global health AI dashboards with community health
storytelling.
Speaker 1 (21:26):
Another hybrid.
Another hybrid.
Can I elaborate?
Yeah, yeah, of course.
Speaker 4 (21:30):
Um so during the pandemic, a lot of Aboriginal
communities in Canada andelsewhere lost a lot of their
elders due to COVID.
And they're in that communityis the history of storytelling
plays a major role.
And it's a passed down of notonly um traditional healing
methods, but also of history andancestors and all.
(21:54):
So losing that completely is aloss to many generations and
several.
So having captured thesestories so that they can be
passed down digitally orotherwise is important.
Speaker 1 (22:09):
Yeah.
So it's it it it has a a wealthof uh of information that we're
losing.
Yeah, that we're losing everytime that uh because uh every
time we lose uh a communityleader, elder that have seen so
much and knows so much and alsohas has learned from his own
elders.
(22:29):
So it's this this kind ofinformation, especially in
communities, indigenous uhpeople's communities, where um
they they have such uh uh a deepunderstanding far beyond that
we can we can grasp of of life,of health, of of um everyday um
information as well,environment, um, agriculture is
(22:52):
is it's limitless theinformation that they have.
So it's it's uh it's importantnot to lose it and not to let it
just slip away.
Speaker 4 (23:00):
And if the digital communications or others can
help capture that, that would bea way to preserve.
Speaker 1 (23:08):
Yeah.
Always with their consent, ofcourse.
Yes, of course.
Then uh next question (23:11):
a universal health algorithm or
millions of micro-local ones.
Speaker 4 (23:21):
Micro-local ones for the context.
Speaker 1 (23:25):
Faster breakthroughs or longer-term resilience?
Speaker 4 (23:31):
Um risk versus benefit.
Okay.
So, for example, during thepandemic, we had to, I was at
Gabby, that was uh you had toget vaccines out fast.
Um, and then you weigh the riskof getting a vaccine out as
soon as possible with versus theshort-term and long-term side
effects.
Um, but these areconsiderations that are taken
(23:55):
into account during the decisionmaking, and regulatory bodies
also help ensure that risks areminimized.
Speaker 1 (24:03):
And I think in that in that aspect, there's no um
easy answer.
There's no easy answer becausethere's there's always gonna be
a lost.
Yes.
And you need to wait which whatwhich one are you willing to to
go ahead and and live with?
Yeah.
So it's no, it's an impossiblesituation, it's an impossible
(24:24):
situation.
Yeah, very tough question.
Yeah, yeah, yeah.
Okay, last flash question.
Could AI train on clinicaltrials or on lived experiences?
Speaker 4 (24:39):
Um both.
I mean the clinical trials.
Yes.
There's no way you can get allof the information during a
clinical trial.
Um the real world evidence oncea drug is on the market in a
not so controlled setting isalso important.
(25:00):
So mm losing any of thatinformation would be tragic.
Speaker 1 (25:05):
Yeah, yeah, it's it's because then when one is
outside, you have the wholeworld that can tell you how it
works.
Speaker 4 (25:13):
And maybe you haven't considered a specific patient
group in your clinical trial.
For example, a lot of pregnantmothers, etc., not to say that
the drug is given to them after,but there are maybe different
minority populations that youhaven't included, whether
racially or disease-specificminorities.
(25:33):
So tracking the information onhow it's affecting these
different groups is veryimportant because we're seeing
in healthcare now, uh afterhaving years and years of data,
we are seeing that drugs mightaffect different minority groups
differently.
So making sure that you try tocapture them in your clinical
(25:57):
trials to begin with, but thenalso after.
Speaker 1 (26:01):
Yeah, there's so many factors.
There's um um the medicinecombination as well.
So a person who has uh anotherunderlying condition that is
taking medicine for somethingelse that interacts with this
new medicine, there's also canbe a reaction for that, also a
genetical predisposition, umeven the the the the health uh
(26:24):
um um uh habits of the person.
How likely is the person to eatwell, how the the food
combination, there even kind offood that can also have a bad
reaction to medicine.
Speaker 3 (26:36):
Yes.
Speaker 1 (26:36):
So it's there's so many factors to take into
account that yeah, not nothingtrumps uh uh real life.
Speaker 4 (26:43):
Agreed, and we recently saw that from the
obesity drugs that have beenidentified where they were
originally not designed forobesity but other diseases, and
then the uh post-marketinformation has shown that they
affect obesity as well, and nowthey've become blockbuster
(27:06):
drugs.
Uh there are a few I shouldn'tshow any biases.
Speaker 1 (27:12):
Yeah, I'm just asking for a friend.
unknown (27:15):
Of course.
Speaker 1 (27:18):
Uh yes, this that's uh and also the way that
interacts with you, uh it can bealso so so different.
Yes.
Um, I have a very low thresholdin any medicine in general
because I I I hate takingmedicine.
So if I can avoid it, I willendure a headache.
Oh wow, stomachache.
I I I prefer to I make a tea, Idrink some water.
(27:39):
Okay.
I I try to not to take a lot ofmedicine, but when I do, I'm
I'm I'm down.
It really is.
It's you're at your last.
Like it's like it's it's umit's crazy because I think it's
because my body's not used tohaving them.
So once in mysticism, it'slike, okay, I'm gonna grab every
single one drop of molecule,and you're gonna get it.
(28:03):
Uh and something funny happenedto to light up the mood.
Um, the first day that I hadwith my husband, uh, we were
supposed to meet uh for for tea.
And I fell asleep because Itook uh um um a pill for my cold
that I had.
I was I had a very, very badcold, and I never take anything.
(28:25):
So I decided to take it, and Ifell completely asleep.
Completely, and because he Ididn't know, but he had a little
bit of anti-allergic into it.
And I also was down.
The poor guy waited for me forover an hour.
Speaker 4 (28:40):
Oh wow, but we met.
Speaker 1 (28:42):
He's a keeper, and he's Swiss, so yeah, being on
top is very important.
So yeah, he really uh uh it'sit's something funny that I say
uh all the time.
Like, I am I'm I'm very much uhI could be a very interesting
specimen.
You would be in this minoritygroup of people.
(29:04):
You would be like, what iswrong with her body that takes
on so many medicines?
Speaker 4 (29:11):
See, this is why we should track post market falling
asleep for an hour.
Speaker 1 (29:15):
There you go, for just a common cold medicine.
Okay, now uh the next the nextgame.
You take the pile there, andthe question um I will ask you
the statement.
You will tell me if it's true,meaning that you believe it's
happening right now and it'sabout to happen, okay, or
(29:37):
futuristic with a tiny robot.
That means that it's somethingsci-fi is gonna happen way in
the future, or it may neverhappen.
Speaker 4 (29:46):
Okay.
Speaker 1 (29:47):
Okay, ready?
Ready.
Health systems will use AI todetect disease outbreaks before
symptoms start to appear.
Patients will have digitalhealth twins that simulate their
(30:08):
care outcomes.
Groups of patients.
Okay, not individual groups,yes.
Okay.
You'll be able to pause yourdigital health footprint.
Speaker 4 (30:23):
Just no.
Neither true or futuristic.
Just no.
No, no pausing.
Speaker 1 (30:29):
No pausing.
Public health policies will beco-drafted by humans and AI.
Governments will use AI topredict well healthcare workers
should be sent before crisishappened.
Diagnosis will include anemotional impact score.
(30:53):
Possible.
Yeah, but it would be nice.
You have the right to deleteyour medical prediction history.
Speaker 4 (31:08):
It it depends on the legislation where you are.
Speaker 1 (31:12):
And what what's what the service uh the purpose is
gonna serve.
Yes.
Because they need informationto treat you.
Speaker 4 (31:18):
And if you don't have it, then yes, but it's still
your information, so you shouldhave a right to it.
Speaker 1 (31:25):
To be forgotten.
Speaker 4 (31:26):
Yeah.
Speaker 1 (31:27):
Yeah.
Tricky, tricky.
Yes, I'll just leave it here.
No question, no answer.
No answer.
AI will be able to detect biasin global health funding models.
unknown (31:39):
True.
Speaker 1 (31:39):
True.
Health insurance will bepartially determined by your
microbiote.
Speaker 4 (31:46):
Depends on legislation.
Speaker 1 (31:49):
So no answer.
Yes.
And lastly, every hospital AIwill have a built-in system that
checks if its decisions arefair and ethical before taking
action.
Speaker 4 (32:03):
So I mean, even between you and I, we might have
a different interpretation ofwhat's fair and ethical.
Speaker 1 (32:09):
Yes.
Speaker 4 (32:09):
So it would be hard to code that into an algorithm.
Speaker 1 (32:13):
So no answer.
No answer.
Love that, love that.
Yeah, neutrality.
Yes.
Speaker 4 (32:19):
And Switzerland.
Speaker 1 (32:20):
Yeah, yeah, we are in Switzerland neutral.
So, final question.
If you could embed oneinvisible value into every
health innovation being builtright now, not a feature, not a
gadget, but a principle.
What would it be and why?
Speaker 4 (32:41):
Um the principle would be if you can identify
both the positives and theharmful behaviors a particular
technology can cause, that youdo everything possible to
prevent the harmful behaviorfrom happening.
Um, so just some examples.
(33:03):
Historically, um during WorldWar II, the first early
computers were used to takecensus information to quickly
identify ethnic minority groupsand round everyone up.
Um with latest AI technologyfor images, it's being used for
(33:26):
lots of interesting things, butit's also being used for child
pornography.
Um so if these harmful outcomescan be identified by the tech
developers before the technologyis released, then putting
precautionary measures in thetechnology, I think, would be a
(33:49):
principle that um would be aresponsibility of the developer.
Um on the other hand, it's notall doom and gloom.
There's also lots of beautifulthings that technology is being
used for, for example, to savethe Amazon forest and identify
deforestation or rehouse 60,000families after the Nepalese
(34:13):
earthquake.
So there's a lot of beautifulthings tech is being used for,
but for the harmful things, justlike Einstein's relativity
theory of the atom that led tothe atomic bomb versus
everything else, I think theyshould be identified, and every
measure that can be taken toprevent the harmful behavior
from happening is theresponsibility of the developer.
Speaker 1 (34:36):
Perfect.
So to put the fences in theright places.
Yes, where possible.
Thank you so much.
Thank you for your time.
Thank you for coming andtalking uh with us.
It's been uh quite a pleasureuh to learn uh about the work
that you do and also uheverything that is um uh the
(34:58):
potential that technology canhave in health.
And thank you for reminding usthat digital health is not about
faster systems or smartertools, it's about people, it's
about justice, it's about askingbetter questions before we
scale solutions.
Because health is not just adata point, it's a human rate.
Speaker 4 (35:21):
Thank you for having me.
It's been a pleasure.
Speaker (35:28):
Thank you for listening to Intangibia, the podcast of
Intangible Law.
Plain talk about intellectualproperty.
Did you like what we talkedtoday?
Please share with your network.
Do you want to learn more aboutintellectual property?
Subscribe now on your favoritepodcast player.
Follow us on Instagram,Facebook, LinkedIn, and Twitter.
(35:49):
Visit our websitewww.intangibilia.com.
Copyright Letista Caminero2020.
All rights reserved.
This podcast is provided forinformation purposes only.
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
The Joe Rogan Experience
The official podcast of comedian Joe Rogan.