August 2, 2025 • 49 mins
In this episode, Jamie Serino and Peter Carucci speak with Dr. Despina Kontos, a professor and cancer researcher at Columbia University, to explore how artificial intelligence is transforming cancer research and patient care. From her roots in engineering to her pioneering work in medical imaging, Dr. Kontos shares how AI is moving from simple computer vision tools to sophisticated predictive analytics that are already saving lives. The conversation dives into the real-world challenges of bringing AI into clinical settings—from legal concerns to integration into existing medical workflows. Dr. Kontos also weighs in on the role of the private sector in driving innovation, the potential of democratizing medical technology to improve access, and why personal responsibility and preventative care are more important than ever. It’s a thought-provoking look at the intersection of technology, medicine, and human decision-making in the era of AI.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Intro (00:02):
Welcome to there's a Lesson in here Somewhere
conversations with interestingpeople with fascinating stories
to tell and from which we candraw important lessons.
Here are your hosts, jamieSerino and Peter Carucci.
Jamie Serino (00:17):
Hello and welcome to.
There's a Lesson in hereSomewhere.
I'm Jamie Serino.
Peter Carucci (00:22):
And I'm Peter Carucci.
Jamie Serino (00:23):
And we're here today with Despina Kontos, a
professor and cancer researcherat Columbia University, and
we're going to talk with herabout her research combining
artificial intelligence andmedical imaging.
It's really fascinating,Despina, welcome.
Despina Kontos (00:39):
It's a pleasure to be here.
Thank you for having me.
Jamie Serino (00:42):
Yeah, thanks for joining us, Despina.
Why don't you start off by justtelling us a little bit more
about yourself?
Despina Kontos (00:49):
Sure.
So, as you said, I'm aprofessor at Columbia University
.
I have a background in computerscience.
You may be wondering what acomputer scientist is doing in a
medical center.
I'm trying to basically combinecomputer science with medicine
and, especially in this time,use artificial initiatives for
(01:11):
the medical center at ColumbiaUniversity more broadly.
And yeah, it's exciting, it'sfun, it's been a great career so
(01:35):
far and been doing this formany years now.
I was at the University ofPennsylvania before joining
Columbia in 2004, and now I'mhere in New York.
Jamie Serino (01:46):
Great.
So you've been sort of at thisfor a little while and in some
ways ahead of the curve, but Iguess in some ways with the
curve, because we heard aboutartificial intelligence and then
we sort of stopped hearingabout it and now of course it's
everywhere.
Yeah, it's everywhere, right,and so maybe you could tell us a
little bit about you know someof the history and you know how
(02:07):
you got into it, what excitedyou about computer science and
then how you got into AI, howyou got into the medical field
as well.
Despina Kontos (02:15):
Sure, sure.
So I grew up sort of like witha screwdriver in my hand, so to
speak.
I was one of those kids wholiked to figure out how
everything works and fixeverything in the house.
And you know, if anything brokein the house my mom would first
ask me to see if I can fix it,and if I couldn't fix it then
she would call the plumber orthe electrician or whoever right
(02:35):
.
So I always had a knack forengineering.
I was also always fascinated bymedicine and biology and I just
always had like a naturalcuriosity to figure out how
things work, you know, in thephysical level and the
mechanical level, like since Iwas sort of very little
Influenced.
I guess both my parents.
(02:55):
My dad was a mathematician,computer scientist, who you know
started I grew up in Greece, acompany there, so a little bit
of an entrepreneur as well.
My mom is a medical doctor andneonatologist.
You know I was alwaysinfluenced by her as well.
But I could never see myselfdealing with little kids as my
profession.
I'd like I am not becoming apediatrician, that is not
(03:17):
happening you know, but I wasequally fascinated by both.
You know, I grew up in thedinner table hearing stories
about, you know, medicine andlittle kids who were in the NICU
and what it took for doctors torevive kids who were, you know,
at danger, and also innovativetechnologies from my dad and
being, you know, naturallyinclined and curious about how
(03:39):
things work.
I was kind of always tornbetween the two and at some
point and you know, when you,yeah, we'll get to the point
that we decide what we actuallywant to do with our lives.
In college, you know, andthings like that, I thought to
myself, well, if I go tomedicine, that's going to take
forever, right, Like I have togo to regular school, then
(04:00):
medical school, then residency,then fellowship Dude.
It's going to be like such along time until I'm out of
school and I have a regular life, Right, and I have like a job
and a salary.
So I said, if I go toengineering, I think life's
going to be quicker.
You know, I can just like plowthrough, be done and be good.
So I went to engineering withthe intent to kind of work with
my dad at the time who washaving a business in Greece
(04:21):
where I grew up, in Greece.
Where I grew up, and one thingled to another and I pursued
sort of graduate work in the US,at which point during my PhD I
had a fellowship and in order tokeep the fellowship and the
financial assistance that cameto it, I had to be involved with
research.
I had to help my professor whoI was doing my PhD with, you
(04:43):
know, in his research lab andstill my plan was to go back and
do the business and the wholeGreek thing and everything.
But it was.
You know, when I started gettinginvolved with research, it was
the sort of first time in mylife that I kind of felt on my
own, without sort of third-partyinfluences that.
You know, this is what I wantto do when I grow up and the
(05:06):
research that my professor wasdoing was related to medical
imaging.
So the fact that it brought inmedicine that I was always
fascinated by and I didn'treally have to go to medical
school to do it.
Well, I could still follow mynatural kind of tendencies in
engineering and my you knownatural curiosities around those
lines.
I really felt like this is whatI want to be when I grow up, so
(05:29):
to speak.
One thing led to another.
You know kind of opportunitiescame up.
I finished my PhD at TempleUniversity in Philadelphia, went
to University of Pennsylvaniato do postdoctoral training you
know here where I was notwanting to go to med school
because I've been in like schoolforever and I kind of maxed out
like everything you know likein terms of being in school.
(05:52):
And and then I joined thefaculty at Penn and started my
career there as an independentinvestigator.
Jamie Serino (06:02):
That's a really cool path.
It was cool that you were ableto combine the two things, or
that you were pursuing one thingand the pull of the medical
piece of it kept drawing you in.
Peter Carucci (06:16):
Yes, yeah, no go ahead, Peter.
No, it's just.
I'm really fascinated by howokay I see how both sides came
together in your formation.
Now, how did we land on AI?
How did that become your focus?
Despina Kontos (06:34):
I studied computer science that's my
undergrad computer engineeringand informatics in Greece.
I did my undergrad there andthen when I came here the PhD
was in computer and informationsystems.
So I was a computer scientistby training and you know there
are different directions onecould take in that field.
There's more of a hardware kindof expertise, there's more of a
(06:55):
you know, all kinds ofdifferent aspects.
But the work that my professorwas doing and I was kind of
attracted to had to do more withsoftware and image processing,
medical image processing, and sowe call it AI now.
But really this field has beenaround for a very long time and
in the field of medical imaging,you know we used to call this
(07:18):
like computer vision, patternrecognition, medical image
analysis.
But a lot of the fundamentalsof what we call now ai used to
be there.
Like my first book aroundneural networks was in fact in
my undergrad in greece and itwas written in greek.
So it's been around for a longtime.
So when you're in thisdiscipline you kind of choose a
direction.
(07:38):
You want to be more hardwarefocused.
So I was thinking like, do Iwant to be making computers, you
know, and things like that,like hands-on, as a computer
scientist.
There was a big direction atthat time was the beginning of
the internet, you know, and allthat stuff was like network.
You know network protocols.
How does you know what gaverise to internet and bandwidth?
Do I want to, you know, be thekind of person who puts together
(08:01):
wires and things about networkarchitecture like man?
not really so I like this was alittle bit more um, at least for
me.
Oh, I could totally see myselfgoing in the hands-on direction
because I really, really lovethe screwdriver.
You know, experience, um, uh,this was kind of that combined,
the sort of more scientific uhpursuit, so to speak.
(08:22):
You know, you know there'sengineering and there's science
and there's an overlap, right,engineering is a little
different than science.
I think this allowed me to alsopursue scientific inquiry in
addition to engineering, and Ialways had that as a person.
I was driven by figuring outhow things work.
Peter Carucci (08:45):
And now you're taking data points right.
As a person, I was driven byfiguring out how things work.
And now you're taking datapoints right systemically and
also from medical imaging, andyou're making predictions, and
so the software you've developedmakes these predictions.
Despina Kontos (08:58):
Yes.
So basically I do a whole bitof different things.
So my fundamental expertise asa computer scientist working in
imaging and medical imaging isto take medical imaging data so
CAT scans, mris, x-rays, youname it everything and develop
(09:20):
the computer science algorithms,the techniques to analyze this
data, to extract informationfrom this data, information
that's not routinely assessed bythe clinician, right, so we can
quantify things, we can measurehow big is your hippocampus you
know in this field, you know meand cancer.
We can identify where is acancer, what type of cancer it
(09:42):
is, how big is the tumor, thevolume, whatever, all kinds of
other properties, and developthe algorithm so that the
computers can do thatautomatically.
That's like my fundamentalbread and butter expertise.
Now, beyond that, what I becamefascinated, what do we do with
this information?
Right, you can calculate allkinds of things from the data.
I'm excited to kind of try topredict what will happen in the
(10:05):
future.
So I'm not as focused about,like the now, the diagnostic
piece.
I mean that we haveradiologists, we have doctors,
they do that pretty well.
I want to be able to likepredict what will happen in the
future and use that informationto better tailor your care now.
So, if you are a healthyindividual, are you at high risk
(10:27):
of developing a certain cancer?
What are the implications ofthat?
You know.
How do you?
What can we do to prevent, toreduce your risk?
What can we?
How should we screen you ifthere's a cancer, that screening
is available?
If there's a cancer thatscreening is not available, can
we use this information to saymaybe this is a good opportunity
to start thinking about how dowe screen these individuals that
are at high risk, for example,ovarian cancer, which is an area
(10:49):
I'm working on right Morerecently, there's no screening
for women for ovarian cancer.
Why is that?
Okay, there are a lot ofreasons why, but can we use some
of these technologies to maybeenable that?
There are some questions, right.
If you have a cancer, how likelyis it to come back?
How likely is it to recur?
What's the timeline?
Based on that information, howaggressive do we need to go with
(11:11):
treatment?
And if you are in treatment,are you responding?
Will this drug help you fullyrespond or not?
And if not, what could be otherdrugs that we could add in the
equation, right?
So I'm interested in predictingthe future and in doing that,
you know we use computeralgorithms and artificial
intelligence now to getinformation from the imaging
(11:33):
data, but also plug in otherdata, because think about it
right, when you go to a doctor,oh, you know.
Good morning Jamie.
What's going on today.
I have this.
I have that.
I have this heart in here.
Whatever I did an x-ray, I didthis.
So the doctor will take thisinformation.
They're going to look at thex-ray.
They're going to ask you aquestion about how you feel
about things.
They're going to look at yourfamily history.
They're going to look at whatmedications you take.
They're going to look how oldyou are.
(12:00):
It's going to be a differentthing if there are a bunch of
comorbidities they have to thinkabout.
So I use artificialintelligence to basically do
that collect all the data, weighthe pros and cons and come up
with a prediction that istailored to your individual
needs.
How do we need to treat jamietoday based on who jamie is
(12:22):
today, versus how do we need totreat Peter today based on who
Peter is, even if they have thesame exact thing?
You know what I mean.
Jamie Serino (12:28):
Yeah, and so right now, a lot of the information
exists in silos.
Yes, there are portals now.
One doctor sees a test fromanother doctor and that's
helping.
But what you're trying to do ishave AI begin to process all
this information, yes, and thenalways have it there.
It could be in the portal or insome report or whatever, and
(12:51):
then better action could betaken.
Despina Kontos (12:54):
Yes, and make it better for them.
Make it easier, you know,because really, in this day and
age, how much time does a doctorhave when you go see them,
right?
Not that much.
So I want to do the uh, I wantto do the laundry for them, have
everything ready, right?
So when you go there, all thisinformation is on their
fingertips and they can bettermake decisions, uh, for you when
(13:16):
you are at the point of care.
That's the goal, you know.
Peter Carucci (13:18):
That's what we aspire to do with this research
you know, I, I my my brain islike coming up with a bazillion
questions for you, but the onethat I'm very fascinated to know
is without you don'tnecessarily have to reveal too
much, but have there beencertain circumstances where this
data has been really, and youruse of AI and the data has been
(13:41):
really instrumental in savinglives?
And are you willing maybe youwant to share a story or two
without going too much?
Despina Kontos (13:49):
I mean, we are still in the.
Peter Carucci (13:51):
Or is it more a general overarching thing that's
happening?
Despina Kontos (13:58):
Yes and no.
You know we're in the earlydays, right.
We have not yet materializedthe impact that could be done
for a host of reasons, notnecessarily because we don't
have the technology for legalimplications, for workflow
implications, for costimplications, for insurance
implications, all kinds ofthings.
I could talk about that forever.
But there are circumstanceswhere AI is helping today to
(14:23):
make things better.
Where AI is helping today tomake things better, for example,
there are tools that aneurosurgeon can use to
visualize the tract of the brainwhen they're doing a surgery,
for example, to extract aglioblastoma or another type of
tumor in the brain.
There's one thing to do theextraction, but the thing is
what's the collateral damagewhen you do that in the brain?
(14:45):
There's one thing to do theextraction, but the thing is
what's the collateral damagewhen you do that?
So, by trying to extract atumor, are you going to you know
damage, like language, you knoware you going to damage other
function based on where thetumor is Cause you have to go in
there and cut things to gettumors out Right.
So you can use AI to takepictures of the brain, you know,
before the surgery, map thetracks of the brain, label them
(15:06):
for their differentfunctionality and live process
during the time of the surgery,guide the surgeon on how to do
the surgery as best as possible,minimizing the potential damage
right.
There's AI that can, for example, in one of the areas I've been
working on is predicting therisk of developing breast cancer
using information from imagessuch as mammographic density and
(15:30):
family history and genetics.
Combining that information andthat can help inform women to
seek screening more frequentscreening, for example, for
high-risk women, supplementalscreening with additional
imaging techniques such as MRIor breast ultrasound and so
forth.
So in radiation treatment, youknow, ai and computer guided
(15:53):
techniques are helping focus theradiation treatment as much as
possible in the area of thetumor where sparing other organs
from the collateral damage ofradiation, because you know, so
there's a lot of applicationsthat are being used.
We can predict which patients inemergency care or in the ICU
may develop sepsis and how tomonitor these patients more
(16:18):
closely.
There's a lot of applicationsand I think more and more
applications will be coming outin that aspect.
Jamie Serino (16:23):
I think more and more applications will be coming
out in that aspect, and so towhat degree are you working also
, then, with the hospital?
Very closely and any of yourwork, is that starting to make
its way into use?
Despina Kontos (16:39):
Absolutely very closely and that's why sometimes
a lot of my colleagues ask me.
Sometimes a lot of mycolleagues ask me I'm a computer
scientist and my employer is aradiology department.
I work for a radiologydepartment.
I'm not in engineering.
I don't teach classes.
I don't teach computer science.
I don't do any of that stuffthat a traditional computer
science professor would do.
(17:04):
So I'm a little bit of a hybridand a bit of an outlier in my
field.
It's exactly because of thatreason, because I want to be in
the source of the data and Iwant to be in daily interaction
with clinicians.
Otherwise my work is, you know,it will keep me going, It'll
give me publications, it willgive me, you know, what
academics need to get promotedand whatever, but like it's not
(17:27):
going to make a difference inanything.
So that's another thing.
I've always been driven as a asan individual, you know, seeing
life a little bit from a morekind of philosophical
perspective and try to findmeaning meaning in my life.
You know, for me making animpact is a very important
aspect of how I find meaning inmy life, and so through my work,
(17:52):
through my, you know, online Idon't like compartmentalizing my
life either very much Like Idon't smell it, it's like, oh,
this is my work life, this is myhome life, this is my this life
, this is my other life.
Peter Carucci (18:01):
You know my life is my home life, this is my this
life, this is my other life.
You know, my life is my life.
Despina Kontos (18:03):
There's one life .
You know I have one life andit's here and now, and my drive
since again this also were someof the things I've always felt
very strongly, since I was a kidhas been to make a difference.
Like you know, when I'm gone,is it important that this being
(18:24):
was around or not.
That's important to me.
Peter Carucci (18:27):
It's a very healthy outlook.
Do you see just changing gearsa hair?
It's strange.
I want to know your predictionsabout what AI will be
predicting for the future.
Ai will be predicting for thefuture.
How much AI will beincorporated into the medical
(18:50):
field compared to where it isright now?
Do you see it like trajectory,going like sky high everywhere,
or is it still a balancing act?
I'm very fascinated.
Despina Kontos (19:04):
In general, in our life, in this society, AI
will take over fully.
It's a tsunami.
It's happening and we haven'teven gotten into quantum
computing yet.
Wait until we get there.
I don't know if our generationwill fully see the fruition of
that, I think our kids for sure.
(19:24):
It's going to be nothing thatyou've ever imagined.
I think A complete revolution,like the industrial revolution.
It's going to be a completelynew era.
Healthcare, I think, is goingto be one of the areas that's
going to be most resistant.
It will know it will penetrate,will take a while.
I'll tell you why.
Because of all the legal,insurance and workflow
(19:47):
considerations, right?
So let's say I have an AI tool,right?
Let's say I have this tool,which I have it.
I make it freely, publiclyavailable for anybody to use it.
You can download it, use it atyour home if you wish, that can
measure your mammographicdensity from a mammogram and can
tell you some you knowinformation about the risk of
developing breast cancer.
Let's say I have it.
Okay.
Let's say I even make it freeavailable.
(20:08):
Let's say we put it in thehospital.
First of all, you have the FDAapproval.
It's all of that.
That landscape is evolving, youknow all of it.
But that's I'm not in thatbusiness.
I'm not in the business of FDAapproval.
So someone has to get it,somebody has to be interested in
, like a company or somebodypicking it up.
You know there's a caveat withFDA that if it is within our
homegrown environment we can putour tools in.
(20:29):
So I'd say, at Columbia I canput it in, but I can't put it at
like other hospitals, right?
Okay, let's say somebody picksit up.
Let's say we solve that problem,which is a big problem.
Who's going to pick it up andactually do that kind of path to
commercialization, translation,fda approvals, all that stuff,
figure that out, ok.
And let's say that the tool iseven fully validated, because
(20:50):
when I'm in the lab I do mydiscovery on a certain
population, a sample size.
You know I don't do it to theuniverse, right?
So if I'm going to use it forthe entire US population or
Europe or Asia, who knows whereit's going to go, you know we
need to do more work to to makesure it works for that data as
well.
So let's say we solve theseproblems.
Let's say we also solve theproblem of a commercial partner
(21:14):
being interested to pick that upand move it forward.
Ok, which are all?
Those two are pretty bigproblems, right?
So let's say we solve them.
Then we get to the hospital.
A health system is a financialinstitution, right?
So there are a ton ofimplications.
So let's say I'm the doctor andthey give me this tool.
This is going to take me moretime to do my job because I have
(21:38):
to do my assessment and then Ihave to check in what the AI
tool says and I have to waitwith my assessment and I have to
figure it out.
So if it takes me five minutesa minute I don't know how much
to read a mammogram, it could bethat this will double the time
that I need to do my job.
Okay, what does it bring in?
Does it reduce the workloadthat I do?
Let's say it says somethingright or wrong.
(22:01):
Who takes the legal liability?
Who's responsible?
Is it the doctor responsible orthe AI tool or the hospital?
How does insurance play out IfI'm a doctor with AI?
Does insurance compensate memore because I use the AI and I
spend more time doing my job andblah, blah, blah, or it doesn't
compensate me.
Does it bring more business inthe hospital or less business?
So there's so many questions.
(22:21):
How much is the license?
If the hospital is going to payX money to get a license for
all its doctors to use my tool,that's a lot of money.
How do they make up for it?
What is the return ofinvestment for them to use these
tools?
And I will say, and I say to allmy colleagues, I still don't
have this answer, and part of meis kind of thankful that I'm
not in that business.
(22:42):
I'm in the business of, like,the scientific discovery.
So I don't, you know, I don'thave to worry as much for my day
job about this, but I worryexistentially, like if all I do
becomes a gimmick on the websitethat somebody can use to have
some fun.
Like, what are we doing here,right?
So?
But I keep asking my colleagueswho are in the healthcare
(23:03):
systems, who are, like you know,cfos, whatever in a health
institution, like have you seenan AI tool that has made a
return on investment for theinstitution that adopted it?
I do not know yet of an AI toolthat has delivered a concrete
ROI.
(23:24):
And I think there's a lot ofeffort right now, all these
companies that are trying tocommercialize these things, to
kind of measure the impact ofthese tools within a healthcare
system, on patients, on thelogistics of healthcare, right,
and so until we work that out, Ithink there's going to be a lot
(23:44):
of resistance from healthcareinstitutions.
The appetite is there, but theimplementation is just tough,
and so eventually there's goingto be a lot of pressure in the
system because non-traditional Ithink healthcare systems are
going to invade this space.
So you know, like Amazon,google, all of these folks like,
(24:06):
they see the value and they'restructuring healthcare models
that are very different than atraditional healthcare system in
a hospital and that is going totake away a lot of care from
the conventional healthcaresystems, I think, because of the
AI.
So they see the value, becausethe way the traditional
healthcare system is built nowit's like a complex, archaic
(24:31):
institution.
To go there and inject thistechnology is so disruptive.
But if you're starting fromscratch and you're going with
this from the beginning, you sayI'm going to use this from the
beginning, I'm going to havefewer doctors, I'm going to
streamline all the patientcommunication through it.
If you build it from thebeginning with that in mind,
that's a whole different story.
Right, you can make it used toyour advantage and you can have,
(24:53):
I think, a substantial returnon investment from the get-go.
And so I think eventuallythere's going to be enough
pressure built up in the systemthat they will have to go that
way to also absorb some of thecost of healthcare that is
increasing.
There's going to be a lot ofchanges right now in healthcare
insurance reimbursements, andhealth systems are going to be
(25:15):
under immense pressure to becomesustainable and viable, and so
I think that's where AI is goingto come in and catalyze a lot
of aspects, and that's wherewe're going to start seeing the
adoption of these tools becomingmore broad.
Jamie Serino (25:35):
Yeah, that was going to be my follow-up was
that what role would the privatesector play?
And it could just be inupending it completely, you know
, democratizing it, so to speak,or, you know, taking an
industry that's ripe for change,for disruption, and yeah, if
they start from scratch, thenthey don't have any of those
issues to deal with.
(25:55):
Going outside of the insurancesystem, you know, came to mind
as well.
So, yeah, a lot of differentscenarios, so it's interesting
that you brought that up.
Peter Carucci (26:05):
It's already begun in a way with, like, a lot
of online virtual care.
Despina Kontos (26:10):
Yes, exactly.
Peter Carucci (26:12):
You know, it's cheaper for them rather than you
know, and the liabilities aredifferent.
Despina Kontos (26:16):
Absolutely.
Peter Carucci (26:17):
And it's interesting you mentioned, like
if there was a company or anorganization that began with
this model in mind, it wouldchange the game.
Despina Kontos (26:29):
Completely.
It would change the game.
Yes, what?
Peter Carucci (26:31):
about the international approach.
I know many countries, like inEurope or around the world,
don't have our for lack of abetter term capitalist,
insurance-based health care.
They have health care.
Do you see any differencebetween, let's say, you know
European countries adopting thiskind of AI or other countries
(26:54):
around the world, or are we theleader here and it has been
coming out of this area?
Despina Kontos (27:02):
I think Europe, specifically from what I hear,
the regulatory framework is evenmore complicated than here, I'm
not seeing it coming out ofEurope.
I don't know about Asia, thoughI think there's a good chance
they could be more innovativethan we are in that sense
because they need it, andthere's also they don't have as
many doctors as we do, and sothey look up to some of these
technologies to help alleviatesome of the problem the fact
(27:27):
that they don't have enoughphysicians to take care of
everybody.
So I'm not sure, but I'm notseeing it coming out of Europe,
in my view.
Peter Carucci (27:38):
I think here will happen eventually.
Despina Kontos (27:41):
Here will happen , you know, within the next like
five to ten years, and then Ithink when that starts happening
, we're going to see some moreadvancements from quantum
computing and then everything weknow will be different.
I mean, that is going to be thereal revolution.
If I was a student now, that'swhat I would be doing my PhD.
Peter Carucci (28:06):
What makes that so special, I guess, compared to
where we are, everything youknow will be different.
Despina Kontos (28:12):
Everything you know, everything we have known
about computers will collapseand change.
So quantum computers are basedon the principles of quantum
physics, and I'm not going to gointo all the details of it.
But basically what quantumphysics is telling us is that
(28:35):
not more than one states arepossible to exist at the same
time between matter and energy.
Regular computers are based onthe binary logic.
You know whatever we do in thecomputers, programming however
you want to call, when you godown, down, down, down, down,
down to the chip level, it's allzero and one.
That's it on or off.
Zero and one.
(28:56):
In your interconnection ofhardware, whatever all the
software, we do everything.
When it boils down, when itbreaks down to the hardware
level, it's all all zero and one.
But that's not going to be thecase anymore, right?
And that is the fundamentallimit of how much computers can
compute, because at the end ofthe day, it's all zero and one.
You can be both zero and one atthe same time, but now at the
(29:20):
hardware level.
But now at the hardware level,you know, zero and one can
coexist at the same time andmultiple calculations can be
happening at the same time.
So you know, things that wouldbe computationally not even
possible could happen in minutesor seconds, you know.
Jamie Serino (29:44):
So this is going to blow up everything as we know
, completely blow up everythingas we know.
Despina Kontos (29:46):
I wish I was around to see it.
Jamie Serino (29:47):
I don't think we'll be around to see it, but I
wish I mean it definitely isdiscussed a lot, but I guess
separating the hype or the sortof future outlook of it versus
the reality of it.
But I know a lot of companiesare preparing for it and working
on it and they keep saying it'sgoing to happen soon.
But then to your point, howlong would it then become more
(30:07):
mainstream?
Et cetera, et cetera.
But then I just read an articlewhere there would be a concern
even about the electrical grid.
Despina Kontos (30:15):
It's going to be the same.
Jamie Serino (30:16):
Do we have the electrical grid to actually
support all that?
Despina Kontos (30:21):
Well, things will change.
Once upon a time, a computerwas an entire room and now it's
in your pocket, right, so wewill evolve.
You know it's going to bedisruptive, everything you know.
It's just.
Jamie Serino (30:33):
That's the nature of human evolution you know, not
the first time, not the lasttime that something like that
happened yeah, and and do yousee, like even some of the
medical technology making itsway into the home?
Like, I definitely see the consand the danger of this, but
could someone you know givethemselves an X-ray or give
themselves some sort of scan?
(30:54):
Right, it's a part of maybe ahousehold piece of equipment or
something.
Despina Kontos (30:59):
I think, so I'm all for it.
You know, don't tell me there'sno danger of being in the
hospital.
You go with one type of thingand you come out with five
others, right.
So there are dangers in both.
Wherever there's human natureinvolved, there are risks,
because, in my view, we're theones who always mess up
everything.
But yeah, I think all thesethings will happen, absolutely.
(31:24):
I'm excited for it.
Jamie Serino (31:27):
Yeah, I mean cause that's really.
It's like democratization.
You see, this access to thingsand you know, down to the
individual level, um is justwhat has been happening over and
over and over again.
Um, so you, you brought up umscanning and trying to predict.
So we talked a little bit aboutusing the AI to sort of draw
(31:49):
all this information together,doing the laundry for the doctor
, you know, as you you weresaying how about, like, scanning
and using, you know, predictivemetrics and doing some sort of
preventative treatment?
Like I do not have a tumortoday, but chances are because
of my family history.
Maybe there are genetic testsalso and then maybe now ai is
(32:12):
picking up something because ofmy age or something they have an
inflammation in my body orwhatever it is.
Um, and then all of a sudden,red flag, you, you might develop
a tumor somewhere.
Like, are you also doing thatsort of predictive?
Despina Kontos (32:26):
Yes, absolutely Absolutely, and I think it's
important to have thisdiscussion, you know, because
through interventions, lifestyleinterventions, you know, I'm
not saying we're going to liketotally prevent cancer or
whatever, but you know theimpact we could have on
healthcare by lifestyleinterventions.
Battling obesity, you know.
You know wellbeing, you know,and so forth, hypertension, you
(32:48):
know all of these factors thatare water, based on lifestyle
and some genetics.
I'm not like saying just alllike lifestyle, but you know, we
can have a you know impact onpublic health.
That is, could have more impactthan you know, for example, all
cancer therapists combined right.
So, and so I think it'simportant to have these
(33:11):
discussions around preventionand and and, for people to be
open to the idea of if, if theywant that in their lives, to
have that information.
Like you know I have a lot ofdiscussions with friends or
colleagues.
I'm like you know, I'm theperson who likes to know, right?
(33:31):
You know I want to do genetictesting.
I want to know what's going onand they're like why do you want
to know if you're going to have, like, alzheimer's?
Well, I want to know because Iwant to be on the lookout.
For example, right Like there'sa whole bunch of therapies now
in development.
You know preventative things.
You know lifestyleinterventions.
If nothing else, I want toprepare myself.
(33:52):
I think it's important to havethat information and have these
conversations.
Peter Carucci (34:01):
You know, as you're saying, that I'm just
thinking about a family member Ihad, who he's since passed, but
he had a certain condition andhe got a scan back and the
doctor said to him hey, I've gotgood news.
This measurement did not go downanymore in the last month, and
(34:22):
so he celebrated by going out tohave the biggest dinner and
eating everything he shouldn'thave eaten in his celebratory
meal.
And I said to him what are youdoing?
The doctor said this isn't inyour diet.
This isn't in your diet.
You're not supposed to eat this.
You got to have more broccoli.
What are you doing?
And he goes hey, I got greatnews.
The prediction came back that Ihaven't lost any more of that
(34:46):
thing.
And this is great and it'sfunny.
You can lead the horse to waterwith a lot of this information,
but you can't make that horsemake the water.
So with humanity, it'sinteresting that AI can make
these predictions and then wemay just go in and mess it right
up.
We might just go in and youknow.
(35:07):
So do you find that?
I mean, do you see that battleplaying out?
Despina Kontos (35:13):
I see it I don't know if I see it as a battle.
You know I'm a huge believer inpersonal choice and personal
responsibility and personalresponsibility.
You know there's a very goodargument of you know the
extending life okay, as aconcept.
(35:33):
You know, I want you to preventthese things and that because I
want you to live longer.
Okay, how much value does thathave to the individual?
Maybe, if some person wants tolive a certain way, it'll be
shorter you know, I respect that, you know or they want to eat
this all day.
They eat all this all day.
Like, what can I tell you?
I'm just going to give you theinformation.
You do what you feel isappropriate for you to do.
I'm not going to tell you, I'mnot going to force you to do
(35:55):
anything.
I'm very against like forcingin that sense.
But I do think it's importantto have the information and
aware, well-informed choices andthen if your strategy is an
exit strategy, that's on youyeah, I, I think you know
(36:15):
there's it, it is, it is on youand, and, I think, different
people.
Jamie Serino (36:21):
I've heard people say, like you know, the mental
health community, that there'sactually an overwhelming amount
of information now about ourhealth.
And you have these sort ofbiohackers, influencers that you
know spend almost every minuteof their day doing something
that is extending their life,you know, or fighting against
your life, you know, and that'sthat's great.
(36:42):
That's one extreme.
But then you know, some peoplethat you know maybe have anxiety
issues are beginning.
You know, am I drinking enoughwater?
Am I exercising enough?
Am I getting enough hours ofsleep?
Am I meditating?
Is my cortisol levels too high?
Are my cortisol levels too high?
And so, yeah, there can be anoverwhelming amount of
information, but I guess then atthat point, each person has to
(37:06):
deal with it and pick and choosewhat they want to incorporate
into their lives.
You know, I guess to that thisis making me think to ask the
question of, like you know, doesthis add, you know, an even
more overwhelming amount ofinformation to the person?
It's just like, hey, theinformation is there, each
individual deals with it, youknow, is that what it is, or is
(37:27):
there some other you know thinghere, anything else that you'd
have to add?
It is like or is there someother you know thing here,
anything else that you'd have toadd?
To some, to something like likethat, that type of future I?
Despina Kontos (37:37):
can only speak for myself and my own experience
.
Right, and at the end of theday, I'm not even a doctor, I'm
just a computer scientist and anengineer.
I personally like theinformation.
I don't feel overwhelmed by theinformation, but I think you
know one of the things we don'tdo very well in our society is
(37:58):
to teach people in oureducational system whatever it's
very like how to use theirbrain, right, we are very kind
of like information focused,like you have to learn, learn,
learn things, but we don'treally teach people how to use
their brain.
In what sense?
Like you know, a lot of whatyou're talking, just talked
(38:19):
about, is like how you sufferfrom your own brain, like how
you can't even untangle your ownthoughts, how you let your
brain drag you down in anydirection, and how we live our
lives in ability and the skillto pause and be the masters of
(38:54):
our own brain, of our ownemotions and, at the end of the
day, in our own destiny.
We are just like being draggedalong, right?
So I think so from myperspective.
So I think so from myperspective.
I like information, I want toknow.
That's always been my thing.
Like I always want to know.
(39:17):
I don't care, I want to know.
I don't care the cost, I wantto know.
But I think it's important to,as a society, to develop the
skills to use our brains and ouremotions in a constructive,
healthy, life-sustaining way.
We don't do that, you know.
(39:37):
Most of what we suffer from isour own thoughts and our own
emotions about something thatdoesn't even exist, like either
something that has happened inthe past or something that we
think that will happen in thefuture.
But we very rarely are presentin the present moment.
Peter Carucci (39:53):
Yeah.
Despina Kontos (39:53):
So, that takes us in a very, completely
different kind of direction thisdiscussion.
But the problem is how?
Is not the information or thetools or whatever is how.
The problem is always us, it'salways it's us.
You know our individualresponse to what is happening.
(40:14):
That is always the problem inour existence, right, I think at
least, and you know our abilityto respond to what is happening
.
That, I think, is where we needto focus as a society and take
control of our ability torespond in a conscious, aware
and health sustaining manner.
That is my view.
Jamie Serino (40:36):
Yeah, it's well said.
I think it involves a littlebit of Taoism, a little bit of
Buddhism.
Yeah, the thought comes in, thethought leaves Right, and you
know it also makes me thinkabout going back to what you
said earlier.
There's a bit of like changemanagement here that needs to
happen as you introduce this newthing, and we talked about
(41:00):
introducing it to the healthsystem, but even introducing it
to doctors, any anyone using,you know, this medical
information, and thenintroducing it even to the
people, the patients and stuff.
So the the change managementpiece is you know how do we use
this information?
Um, either as a practitioner oras the patient receiving it.
(41:22):
Um, and that change managementpiece, you know that that that's
where like disruption comes in,and disruption is successful
when the change management partgoes well.
Um, when it doesn't go well,that's where like disruption
comes in and disruption issuccessful when the change
management part goes well.
When it doesn't go well, that'swhen it gets all kind of clunky
.
So I mean, do you ever havethose types of discussions in
terms of like, okay, now we'reintroducing this new thing, you
(41:43):
know what's going to be theresistance to it.
How do we get over thatresistance?
That's the change managementand organizational, that's the
behavioral Exactly.
Despina Kontos (41:51):
Exactly, we have these discussions all the time
and again, as I said, I'm justthe engineer, guys.
I am making the gadgets andbringing them to the doorstep.
That is my expertise, mytraining, my job, my calling, my
thing.
I make the gadgets and bringthem to your doorstep.
But I think the process ofevaluating whether or not
(42:11):
something should actually beused by a physician should
actually be communicated to thepatients.
That's a whole differentscientific field.
Like there is what we call likeimplementation science,
comparative effectivenessscience, outcomes research.
There are actual experts aboutthis and we kind of bypass that
whole thing.
(42:32):
We bypass that thing.
We just take it from me, theengineer what do I know?
And try to like shove it in theface of the health system with
the hands of the physician.
We bypass the whole process,folks.
Like we need to engage theseexperts to do their job right
and we need to work together.
Like I'll give you the gadget.
You take this gadget, see if itmakes sense or not.
(42:53):
If you have some feedback, giveit back to me.
I'll tweak it, I'll change it,I'll do whatever it takes you
know, and then the health systemneeds to be communications Like
.
there needs to be anunderstanding that, at the end
of the day, we need to do thisto better our society and
healthcare and health patients.
At the end of the day, we needto do this to better our society
and health care and healthpatients at the end of the day,
(43:14):
right.
If this is not our goal, wewill fail as a society in this,
and so we need to have thesekind of open communications and
coordination between people likeme.
Peter Carucci (43:36):
And coordination between people like me, people
who are in the middle and peoplewho are on the other side, to
be able to successfully bringthis to help us, to help us as
humanity right.
And you're currently using, orrather you created, this kind of
system and you're working withAI to make certain predictions
about specifically certaincancers and whatnot.
About specifically certaincancers and whatnot, and I
recall in an earlierconversation we had, you're also
looking at it in terms of like,maybe predicting Alzheimer's.
Despina Kontos (43:55):
Yeah, not me, my colleagues, right, I work in
cancer specifically, butcolleagues of mine.
Peter Carucci (44:01):
I guess this is a field kind of the predictive
Generally, overall, Exactly, I'dlove to hear a little more
about what you see that futurelooking like, or what that looks
like.
Despina Kontos (44:12):
I think there are similar considerations
across the board.
If it's cancer, if it'sAlzheimer's, if it's cardiac
disease, there's similar trendsand similar work happening
across the board.
And we have the data, thetechnology, the expertise to
develop these technologies andthese tools that can help us
(44:32):
personalize care.
And so, again, the questionbecomes what are the barriers,
the opportunities?
And we need to decide as asociety, how do we want to use
these technologies?
I think it's up to us to decidecollectively what we want to do
with it.
Peter Carucci (44:50):
Do you really think it's up to us to decide
collectively what we want to dowith it?
Do you really think it's up tous to decide, or do you think
it's?
Despina Kontos (44:59):
the financial impact.
I know Life is complicated andthere's all kinds of interests,
guys, but again, from myperspective, if we don't all
take personal responsibility, weare doomed.
It's very easy to say it'ssomebody else's fault.
That doesn't help anybody.
It's our collective, individualresponsibility that's
responsible for everything.
We vote, we elect the peoplewho govern us.
(45:22):
We buy products.
We don't buy products.
We invest this whole thing.
I've never been a person wholike, oh, it's like the.
They right, they, they want todo this.
They, who are they?
It's us.
I don't know this.
They.
I've never seen this they.
Have you seen this?
They?
So I think it's.
We all have to take personalresponsibility.
(45:42):
There's no other way out.
Peter Carucci (45:44):
But in this polemic we're discussing the
possibility of, let's say Idon't want to say names but a
large company Amazon or Googleor any large company could just
jump the shark here, take thistechnology and try to force its
hand into healthcare for thefinancial benefits.
(46:05):
Are you the end?
Despina Kontos (46:06):
user of this.
They all depend on us.
They all depend on the end user?
Peter Carucci (46:09):
Who are they depending on?
Who are they going to sell itto?
Despina Kontos (46:11):
It's your money and my money, and you know who's
going to pay for this right,who's going to pay for it.
So we don't diminish the roleof personal responsibility.
It is the only way out.
The only way out, I think Allright?
Jamie Serino (46:29):
Well, Despina, as we look to wrap up here, is
there anything that we didn'task or anything that you would
want to make sure to add?
Despina Kontos (46:39):
You know, I think the future is bright.
If we want to make it bright.
The future can be very doom andgloomy if we want to make it
doom and gloomy.
History has shown that, youknow, us humans are probably the
worst kind on the planet.
But it has also shown that weare probably the worst kind on
the planet, but it has alsoshown that we're probably the
best kind on the planet as well.
So I think it's up to us todecide how we want to flip that
coin.
Yeah, and, of course, everybodyin their daily lives to take
(47:02):
personal responsibility foreverything that they do, as
little as or big, whateverthey're doing, every moment
counts, I think in this game.
Jamie Serino (47:13):
Yeah, and that's a really good message because,
you know, of course we didn'treally touch upon this, but
there's always these likeend-of-the-world scenarios with
AI, terminator, you know,missiles launching.
And it's interesting because,you know, we talked about Asia
and the Asian cultures tend notto have that mythology and the
(47:34):
Western cultures tend to havethat mythology that the robots
are going to rise up and take usover, and you know, and you
know, I don't know who's rightthere, but you're sort of saying
that the individuals have theresponsibility and they have the
wherewithal and they have thecontrol.
Despina Kontos (47:48):
Really, in the end, Show me a time in human
history where robots were nottaking over Quote robots when
there was peace and prosperityand that everybody was in La La
Land.
You know human history is, youknow, continuously in turmoil,
mayhem, wars, this disaster Like.
(48:11):
Show me a time in human historywhere this does not exist.
So I think there's a little bitof a hype there and, at the end
of the day, if the robots takeover and destroy us, we'll
probably deserve it, you know,because we let them do it.
Jamie Serino (48:23):
We were asking for it.
Despina Kontos (48:25):
You know, I also have a perspective.
Why are we always doom andgloom Like the robots are going
to be like the worst of us.
There's a fair chance that thatmay be the best of us, you know
the best?
yeah, I think we're totallycapable of destroying this
planet without any help from therobots.
We don't need the robots toself-destruct.
I think we're doing a good jobat it as it is.
(48:46):
So, who knows, maybe the robotsare going to save us because
they might tap into the best ofour intelligence.
That's another flip side of thestory.
That's a scenario.
Jamie Serino (48:57):
I like that A message of control, control and
Despina.
It's been fascinating.
We learned so much here.
We're probably gonna have tocome back and talk to you some
more because we still have, Ithink, a thousand questions, but
I want to thank you.
I want to thank everyone forwatching and listening and we
will see you all next time.
Despina Kontos (49:17):
Thank you.
Thank you for having me andlet's get together again before
the robots come.
Okay, thank you All right.
Jamie Serino (49:24):
Thank you.
Welcome to there's a Lesson in here Somewhere
conversations with interestingpeople with fascinating stories
to tell and from which we candraw important lessons.
Here are your hosts, jamieSerino and Peter Carucci.
Jamie Serino (00:17):
Hello and welcome to.
There's a Lesson in hereSomewhere.
I'm Jamie Serino.
Peter Carucci (00:22):
And I'm Peter Carucci.
Jamie Serino (00:23):
And we're here today with Despina Kontos, a
professor and cancer researcherat Columbia University, and
we're going to talk with herabout her research combining
artificial intelligence andmedical imaging.
It's really fascinating,Despina, welcome.
Despina Kontos (00:39):
It's a pleasure to be here.
Thank you for having me.
Jamie Serino (00:42):
Yeah, thanks for joining us, Despina.
Why don't you start off by justtelling us a little bit more
about yourself?
Despina Kontos (00:49):
Sure.
So, as you said, I'm aprofessor at Columbia University
.
I have a background in computerscience.
You may be wondering what acomputer scientist is doing in a
medical center.
I'm trying to basically combinecomputer science with medicine
and, especially in this time,use artificial initiatives for
(01:11):
the medical center at ColumbiaUniversity more broadly.
And yeah, it's exciting, it'sfun, it's been a great career so
(01:35):
far and been doing this formany years now.
I was at the University ofPennsylvania before joining
Columbia in 2004, and now I'mhere in New York.
Jamie Serino (01:46):
Great.
So you've been sort of at thisfor a little while and in some
ways ahead of the curve, but Iguess in some ways with the
curve, because we heard aboutartificial intelligence and then
we sort of stopped hearingabout it and now of course it's
everywhere.
Yeah, it's everywhere, right,and so maybe you could tell us a
little bit about you know someof the history and you know how
(02:07):
you got into it, what excitedyou about computer science and
then how you got into AI, howyou got into the medical field
as well.
Despina Kontos (02:15):
Sure, sure.
So I grew up sort of like witha screwdriver in my hand, so to
speak.
I was one of those kids wholiked to figure out how
everything works and fixeverything in the house.
And you know, if anything brokein the house my mom would first
ask me to see if I can fix it,and if I couldn't fix it then
she would call the plumber orthe electrician or whoever right
(02:35):
.
So I always had a knack forengineering.
I was also always fascinated bymedicine and biology and I just
always had like a naturalcuriosity to figure out how
things work, you know, in thephysical level and the
mechanical level, like since Iwas sort of very little
Influenced.
I guess both my parents.
(02:55):
My dad was a mathematician,computer scientist, who you know
started I grew up in Greece, acompany there, so a little bit
of an entrepreneur as well.
My mom is a medical doctor andneonatologist.
You know I was alwaysinfluenced by her as well.
But I could never see myselfdealing with little kids as my
profession.
I'd like I am not becoming apediatrician, that is not
(03:17):
happening you know, but I wasequally fascinated by both.
You know, I grew up in thedinner table hearing stories
about, you know, medicine andlittle kids who were in the NICU
and what it took for doctors torevive kids who were, you know,
at danger, and also innovativetechnologies from my dad and
being, you know, naturallyinclined and curious about how
(03:39):
things work.
I was kind of always tornbetween the two and at some
point and you know, when you,yeah, we'll get to the point
that we decide what we actuallywant to do with our lives.
In college, you know, andthings like that, I thought to
myself, well, if I go tomedicine, that's going to take
forever, right, Like I have togo to regular school, then
(04:00):
medical school, then residency,then fellowship Dude.
It's going to be like such along time until I'm out of
school and I have a regular life, Right, and I have like a job
and a salary.
So I said, if I go toengineering, I think life's
going to be quicker.
You know, I can just like plowthrough, be done and be good.
So I went to engineering withthe intent to kind of work with
my dad at the time who washaving a business in Greece
(04:21):
where I grew up, in Greece.
Where I grew up, and one thingled to another and I pursued
sort of graduate work in the US,at which point during my PhD I
had a fellowship and in order tokeep the fellowship and the
financial assistance that cameto it, I had to be involved with
research.
I had to help my professor whoI was doing my PhD with, you
(04:43):
know, in his research lab andstill my plan was to go back and
do the business and the wholeGreek thing and everything.
But it was.
You know, when I started gettinginvolved with research, it was
the sort of first time in mylife that I kind of felt on my
own, without sort of third-partyinfluences that.
You know, this is what I wantto do when I grow up and the
(05:06):
research that my professor wasdoing was related to medical
imaging.
So the fact that it brought inmedicine that I was always
fascinated by and I didn'treally have to go to medical
school to do it.
Well, I could still follow mynatural kind of tendencies in
engineering and my you knownatural curiosities around those
lines.
I really felt like this is whatI want to be when I grow up, so
(05:29):
to speak.
One thing led to another.
You know kind of opportunitiescame up.
I finished my PhD at TempleUniversity in Philadelphia, went
to University of Pennsylvaniato do postdoctoral training you
know here where I was notwanting to go to med school
because I've been in like schoolforever and I kind of maxed out
like everything you know likein terms of being in school.
(05:52):
And and then I joined thefaculty at Penn and started my
career there as an independentinvestigator.
Jamie Serino (06:02):
That's a really cool path.
It was cool that you were ableto combine the two things, or
that you were pursuing one thingand the pull of the medical
piece of it kept drawing you in.
Peter Carucci (06:16):
Yes, yeah, no go ahead, Peter.
No, it's just.
I'm really fascinated by howokay I see how both sides came
together in your formation.
Now, how did we land on AI?
How did that become your focus?
Despina Kontos (06:34):
I studied computer science that's my
undergrad computer engineeringand informatics in Greece.
I did my undergrad there andthen when I came here the PhD
was in computer and informationsystems.
So I was a computer scientistby training and you know there
are different directions onecould take in that field.
There's more of a hardware kindof expertise, there's more of a
(06:55):
you know, all kinds ofdifferent aspects.
But the work that my professorwas doing and I was kind of
attracted to had to do more withsoftware and image processing,
medical image processing, and sowe call it AI now.
But really this field has beenaround for a very long time and
in the field of medical imaging,you know we used to call this
(07:18):
like computer vision, patternrecognition, medical image
analysis.
But a lot of the fundamentalsof what we call now ai used to
be there.
Like my first book aroundneural networks was in fact in
my undergrad in greece and itwas written in greek.
So it's been around for a longtime.
So when you're in thisdiscipline you kind of choose a
direction.
(07:38):
You want to be more hardwarefocused.
So I was thinking like, do Iwant to be making computers, you
know, and things like that,like hands-on, as a computer
scientist.
There was a big direction atthat time was the beginning of
the internet, you know, and allthat stuff was like network.
You know network protocols.
How does you know what gaverise to internet and bandwidth?
Do I want to, you know, be thekind of person who puts together
(08:01):
wires and things about networkarchitecture like man?
not really so I like this was alittle bit more um, at least for
me.
Oh, I could totally see myselfgoing in the hands-on direction
because I really, really lovethe screwdriver.
You know, experience, um, uh,this was kind of that combined,
the sort of more scientific uhpursuit, so to speak.
(08:22):
You know, you know there'sengineering and there's science
and there's an overlap, right,engineering is a little
different than science.
I think this allowed me to alsopursue scientific inquiry in
addition to engineering, and Ialways had that as a person.
I was driven by figuring outhow things work.
Peter Carucci (08:45):
And now you're taking data points right.
As a person, I was driven byfiguring out how things work.
And now you're taking datapoints right systemically and
also from medical imaging, andyou're making predictions, and
so the software you've developedmakes these predictions.
Despina Kontos (08:58):
Yes.
So basically I do a whole bitof different things.
So my fundamental expertise asa computer scientist working in
imaging and medical imaging isto take medical imaging data so
CAT scans, mris, x-rays, youname it everything and develop
(09:20):
the computer science algorithms,the techniques to analyze this
data, to extract informationfrom this data, information
that's not routinely assessed bythe clinician, right, so we can
quantify things, we can measurehow big is your hippocampus you
know in this field, you know meand cancer.
We can identify where is acancer, what type of cancer it
(09:42):
is, how big is the tumor, thevolume, whatever, all kinds of
other properties, and developthe algorithm so that the
computers can do thatautomatically.
That's like my fundamentalbread and butter expertise.
Now, beyond that, what I becamefascinated, what do we do with
this information?
Right, you can calculate allkinds of things from the data.
I'm excited to kind of try topredict what will happen in the
(10:05):
future.
So I'm not as focused about,like the now, the diagnostic
piece.
I mean that we haveradiologists, we have doctors,
they do that pretty well.
I want to be able to likepredict what will happen in the
future and use that informationto better tailor your care now.
So, if you are a healthyindividual, are you at high risk
(10:27):
of developing a certain cancer?
What are the implications ofthat?
You know.
How do you?
What can we do to prevent, toreduce your risk?
What can we?
How should we screen you ifthere's a cancer, that screening
is available?
If there's a cancer thatscreening is not available, can
we use this information to saymaybe this is a good opportunity
to start thinking about how dowe screen these individuals that
are at high risk, for example,ovarian cancer, which is an area
(10:49):
I'm working on right Morerecently, there's no screening
for women for ovarian cancer.
Why is that?
Okay, there are a lot ofreasons why, but can we use some
of these technologies to maybeenable that?
There are some questions, right.
If you have a cancer, how likelyis it to come back?
How likely is it to recur?
What's the timeline?
Based on that information, howaggressive do we need to go with
(11:11):
treatment?
And if you are in treatment,are you responding?
Will this drug help you fullyrespond or not?
And if not, what could be otherdrugs that we could add in the
equation, right?
So I'm interested in predictingthe future and in doing that,
you know we use computeralgorithms and artificial
intelligence now to getinformation from the imaging
(11:33):
data, but also plug in otherdata, because think about it
right, when you go to a doctor,oh, you know.
Good morning Jamie.
What's going on today.
I have this.
I have that.
I have this heart in here.
Whatever I did an x-ray, I didthis.
So the doctor will take thisinformation.
They're going to look at thex-ray.
They're going to ask you aquestion about how you feel
about things.
They're going to look at yourfamily history.
They're going to look at whatmedications you take.
They're going to look how oldyou are.
(12:00):
It's going to be a differentthing if there are a bunch of
comorbidities they have to thinkabout.
So I use artificialintelligence to basically do
that collect all the data, weighthe pros and cons and come up
with a prediction that istailored to your individual
needs.
How do we need to treat jamietoday based on who jamie is
(12:22):
today, versus how do we need totreat Peter today based on who
Peter is, even if they have thesame exact thing?
You know what I mean.
Jamie Serino (12:28):
Yeah, and so right now, a lot of the information
exists in silos.
Yes, there are portals now.
One doctor sees a test fromanother doctor and that's
helping.
But what you're trying to do ishave AI begin to process all
this information, yes, and thenalways have it there.
It could be in the portal or insome report or whatever, and
(12:51):
then better action could betaken.
Despina Kontos (12:54):
Yes, and make it better for them.
Make it easier, you know,because really, in this day and
age, how much time does a doctorhave when you go see them,
right?
Not that much.
So I want to do the uh, I wantto do the laundry for them, have
everything ready, right?
So when you go there, all thisinformation is on their
fingertips and they can bettermake decisions, uh, for you when
(13:16):
you are at the point of care.
That's the goal, you know.
Peter Carucci (13:18):
That's what we aspire to do with this research
you know, I, I my my brain islike coming up with a bazillion
questions for you, but the onethat I'm very fascinated to know
is without you don'tnecessarily have to reveal too
much, but have there beencertain circumstances where this
data has been really, and youruse of AI and the data has been
(13:41):
really instrumental in savinglives?
And are you willing maybe youwant to share a story or two
without going too much?
Despina Kontos (13:49):
I mean, we are still in the.
Peter Carucci (13:51):
Or is it more a general overarching thing that's
happening?
Despina Kontos (13:58):
Yes and no.
You know we're in the earlydays, right.
We have not yet materializedthe impact that could be done
for a host of reasons, notnecessarily because we don't
have the technology for legalimplications, for workflow
implications, for costimplications, for insurance
implications, all kinds ofthings.
I could talk about that forever.
But there are circumstanceswhere AI is helping today to
(14:23):
make things better.
Where AI is helping today tomake things better, for example,
there are tools that aneurosurgeon can use to
visualize the tract of the brainwhen they're doing a surgery,
for example, to extract aglioblastoma or another type of
tumor in the brain.
There's one thing to do theextraction, but the thing is
what's the collateral damagewhen you do that in the brain?
(14:45):
There's one thing to do theextraction, but the thing is
what's the collateral damagewhen you do that?
So, by trying to extract atumor, are you going to you know
damage, like language, you knoware you going to damage other
function based on where thetumor is Cause you have to go in
there and cut things to gettumors out Right.
So you can use AI to takepictures of the brain, you know,
before the surgery, map thetracks of the brain, label them
(15:06):
for their differentfunctionality and live process
during the time of the surgery,guide the surgeon on how to do
the surgery as best as possible,minimizing the potential damage
right.
There's AI that can, for example, in one of the areas I've been
working on is predicting therisk of developing breast cancer
using information from imagessuch as mammographic density and
(15:30):
family history and genetics.
Combining that information andthat can help inform women to
seek screening more frequentscreening, for example, for
high-risk women, supplementalscreening with additional
imaging techniques such as MRIor breast ultrasound and so
forth.
So in radiation treatment, youknow, ai and computer guided
(15:53):
techniques are helping focus theradiation treatment as much as
possible in the area of thetumor where sparing other organs
from the collateral damage ofradiation, because you know, so
there's a lot of applicationsthat are being used.
We can predict which patients inemergency care or in the ICU
may develop sepsis and how tomonitor these patients more
(16:18):
closely.
There's a lot of applicationsand I think more and more
applications will be coming outin that aspect.
Jamie Serino (16:23):
I think more and more applications will be coming
out in that aspect, and so towhat degree are you working also
, then, with the hospital?
Very closely and any of yourwork, is that starting to make
its way into use?
Despina Kontos (16:39):
Absolutely very closely and that's why sometimes
a lot of my colleagues ask me.
Sometimes a lot of mycolleagues ask me I'm a computer
scientist and my employer is aradiology department.
I work for a radiologydepartment.
I'm not in engineering.
I don't teach classes.
I don't teach computer science.
I don't do any of that stuffthat a traditional computer
science professor would do.
(17:04):
So I'm a little bit of a hybridand a bit of an outlier in my
field.
It's exactly because of thatreason, because I want to be in
the source of the data and Iwant to be in daily interaction
with clinicians.
Otherwise my work is, you know,it will keep me going, It'll
give me publications, it willgive me, you know, what
academics need to get promotedand whatever, but like it's not
(17:27):
going to make a difference inanything.
So that's another thing.
I've always been driven as a asan individual, you know, seeing
life a little bit from a morekind of philosophical
perspective and try to findmeaning meaning in my life.
You know, for me making animpact is a very important
aspect of how I find meaning inmy life, and so through my work,
(17:52):
through my, you know, online Idon't like compartmentalizing my
life either very much Like Idon't smell it, it's like, oh,
this is my work life, this is myhome life, this is my this life
, this is my other life.
Peter Carucci (18:01):
You know my life is my home life, this is my this
life, this is my other life.
You know, my life is my life.
Despina Kontos (18:03):
There's one life .
You know I have one life andit's here and now, and my drive
since again this also were someof the things I've always felt
very strongly, since I was a kidhas been to make a difference.
Like you know, when I'm gone,is it important that this being
(18:24):
was around or not.
That's important to me.
Peter Carucci (18:27):
It's a very healthy outlook.
Do you see just changing gearsa hair?
It's strange.
I want to know your predictionsabout what AI will be
predicting for the future.
Ai will be predicting for thefuture.
How much AI will beincorporated into the medical
(18:50):
field compared to where it isright now?
Do you see it like trajectory,going like sky high everywhere,
or is it still a balancing act?
I'm very fascinated.
Despina Kontos (19:04):
In general, in our life, in this society, AI
will take over fully.
It's a tsunami.
It's happening and we haven'teven gotten into quantum
computing yet.
Wait until we get there.
I don't know if our generationwill fully see the fruition of
that, I think our kids for sure.
(19:24):
It's going to be nothing thatyou've ever imagined.
I think A complete revolution,like the industrial revolution.
It's going to be a completelynew era.
Healthcare, I think, is goingto be one of the areas that's
going to be most resistant.
It will know it will penetrate,will take a while.
I'll tell you why.
Because of all the legal,insurance and workflow
(19:47):
considerations, right?
So let's say I have an AI tool,right?
Let's say I have this tool,which I have it.
I make it freely, publiclyavailable for anybody to use it.
You can download it, use it atyour home if you wish, that can
measure your mammographicdensity from a mammogram and can
tell you some you knowinformation about the risk of
developing breast cancer.
Let's say I have it.
Okay.
Let's say I even make it freeavailable.
(20:08):
Let's say we put it in thehospital.
First of all, you have the FDAapproval.
It's all of that.
That landscape is evolving, youknow all of it.
But that's I'm not in thatbusiness.
I'm not in the business of FDAapproval.
So someone has to get it,somebody has to be interested in
, like a company or somebodypicking it up.
You know there's a caveat withFDA that if it is within our
homegrown environment we can putour tools in.
(20:29):
So I'd say, at Columbia I canput it in, but I can't put it at
like other hospitals, right?
Okay, let's say somebody picksit up.
Let's say we solve that problem,which is a big problem.
Who's going to pick it up andactually do that kind of path to
commercialization, translation,fda approvals, all that stuff,
figure that out, ok.
And let's say that the tool iseven fully validated, because
(20:50):
when I'm in the lab I do mydiscovery on a certain
population, a sample size.
You know I don't do it to theuniverse, right?
So if I'm going to use it forthe entire US population or
Europe or Asia, who knows whereit's going to go, you know we
need to do more work to to makesure it works for that data as
well.
So let's say we solve theseproblems.
Let's say we also solve theproblem of a commercial partner
(21:14):
being interested to pick that upand move it forward.
Ok, which are all?
Those two are pretty bigproblems, right?
So let's say we solve them.
Then we get to the hospital.
A health system is a financialinstitution, right?
So there are a ton ofimplications.
So let's say I'm the doctor andthey give me this tool.
This is going to take me moretime to do my job because I have
(21:38):
to do my assessment and then Ihave to check in what the AI
tool says and I have to waitwith my assessment and I have to
figure it out.
So if it takes me five minutesa minute I don't know how much
to read a mammogram, it could bethat this will double the time
that I need to do my job.
Okay, what does it bring in?
Does it reduce the workloadthat I do?
Let's say it says somethingright or wrong.
(22:01):
Who takes the legal liability?
Who's responsible?
Is it the doctor responsible orthe AI tool or the hospital?
How does insurance play out IfI'm a doctor with AI?
Does insurance compensate memore because I use the AI and I
spend more time doing my job andblah, blah, blah, or it doesn't
compensate me.
Does it bring more business inthe hospital or less business?
So there's so many questions.
(22:21):
How much is the license?
If the hospital is going to payX money to get a license for
all its doctors to use my tool,that's a lot of money.
How do they make up for it?
What is the return ofinvestment for them to use these
tools?
And I will say, and I say to allmy colleagues, I still don't
have this answer, and part of meis kind of thankful that I'm
not in that business.
(22:42):
I'm in the business of, like,the scientific discovery.
So I don't, you know, I don'thave to worry as much for my day
job about this, but I worryexistentially, like if all I do
becomes a gimmick on the websitethat somebody can use to have
some fun.
Like, what are we doing here,right?
So?
But I keep asking my colleagueswho are in the healthcare
(23:03):
systems, who are, like you know,cfos, whatever in a health
institution, like have you seenan AI tool that has made a
return on investment for theinstitution that adopted it?
I do not know yet of an AI toolthat has delivered a concrete
ROI.
(23:24):
And I think there's a lot ofeffort right now, all these
companies that are trying tocommercialize these things, to
kind of measure the impact ofthese tools within a healthcare
system, on patients, on thelogistics of healthcare, right,
and so until we work that out, Ithink there's going to be a lot
(23:44):
of resistance from healthcareinstitutions.
The appetite is there, but theimplementation is just tough,
and so eventually there's goingto be a lot of pressure in the
system because non-traditional Ithink healthcare systems are
going to invade this space.
So you know, like Amazon,google, all of these folks like,
(24:06):
they see the value and they'restructuring healthcare models
that are very different than atraditional healthcare system in
a hospital and that is going totake away a lot of care from
the conventional healthcaresystems, I think, because of the
AI.
So they see the value, becausethe way the traditional
healthcare system is built nowit's like a complex, archaic
(24:31):
institution.
To go there and inject thistechnology is so disruptive.
But if you're starting fromscratch and you're going with
this from the beginning, you sayI'm going to use this from the
beginning, I'm going to havefewer doctors, I'm going to
streamline all the patientcommunication through it.
If you build it from thebeginning with that in mind,
that's a whole different story.
Right, you can make it used toyour advantage and you can have,
(24:53):
I think, a substantial returnon investment from the get-go.
And so I think eventuallythere's going to be enough
pressure built up in the systemthat they will have to go that
way to also absorb some of thecost of healthcare that is
increasing.
There's going to be a lot ofchanges right now in healthcare
insurance reimbursements, andhealth systems are going to be
(25:15):
under immense pressure to becomesustainable and viable, and so
I think that's where AI is goingto come in and catalyze a lot
of aspects, and that's wherewe're going to start seeing the
adoption of these tools becomingmore broad.
Jamie Serino (25:35):
Yeah, that was going to be my follow-up was
that what role would the privatesector play?
And it could just be inupending it completely, you know
, democratizing it, so to speak,or, you know, taking an
industry that's ripe for change,for disruption, and yeah, if
they start from scratch, thenthey don't have any of those
issues to deal with.
(25:55):
Going outside of the insurancesystem, you know, came to mind
as well.
So, yeah, a lot of differentscenarios, so it's interesting
that you brought that up.
Peter Carucci (26:05):
It's already begun in a way with, like, a lot
of online virtual care.
Despina Kontos (26:10):
Yes, exactly.
Peter Carucci (26:12):
You know, it's cheaper for them rather than you
know, and the liabilities aredifferent.
Despina Kontos (26:16):
Absolutely.
Peter Carucci (26:17):
And it's interesting you mentioned, like
if there was a company or anorganization that began with
this model in mind, it wouldchange the game.
Despina Kontos (26:29):
Completely.
It would change the game.
Yes, what?
Peter Carucci (26:31):
about the international approach.
I know many countries, like inEurope or around the world,
don't have our for lack of abetter term capitalist,
insurance-based health care.
They have health care.
Do you see any differencebetween, let's say, you know
European countries adopting thiskind of AI or other countries
(26:54):
around the world, or are we theleader here and it has been
coming out of this area?
Despina Kontos (27:02):
I think Europe, specifically from what I hear,
the regulatory framework is evenmore complicated than here, I'm
not seeing it coming out ofEurope.
I don't know about Asia, thoughI think there's a good chance
they could be more innovativethan we are in that sense
because they need it, andthere's also they don't have as
many doctors as we do, and sothey look up to some of these
technologies to help alleviatesome of the problem the fact
(27:27):
that they don't have enoughphysicians to take care of
everybody.
So I'm not sure, but I'm notseeing it coming out of Europe,
in my view.
Peter Carucci (27:38):
I think here will happen eventually.
Despina Kontos (27:41):
Here will happen , you know, within the next like
five to ten years, and then Ithink when that starts happening
, we're going to see some moreadvancements from quantum
computing and then everything weknow will be different.
I mean, that is going to be thereal revolution.
If I was a student now, that'swhat I would be doing my PhD.
Peter Carucci (28:06):
What makes that so special, I guess, compared to
where we are, everything youknow will be different.
Despina Kontos (28:12):
Everything you know, everything we have known
about computers will collapseand change.
So quantum computers are basedon the principles of quantum
physics, and I'm not going to gointo all the details of it.
But basically what quantumphysics is telling us is that
(28:35):
not more than one states arepossible to exist at the same
time between matter and energy.
Regular computers are based onthe binary logic.
You know whatever we do in thecomputers, programming however
you want to call, when you godown, down, down, down, down,
down to the chip level, it's allzero and one.
That's it on or off.
Zero and one.
(28:56):
In your interconnection ofhardware, whatever all the
software, we do everything.
When it boils down, when itbreaks down to the hardware
level, it's all all zero and one.
But that's not going to be thecase anymore, right?
And that is the fundamentallimit of how much computers can
compute, because at the end ofthe day, it's all zero and one.
You can be both zero and one atthe same time, but now at the
(29:20):
hardware level.
But now at the hardware level,you know, zero and one can
coexist at the same time andmultiple calculations can be
happening at the same time.
So you know, things that wouldbe computationally not even
possible could happen in minutesor seconds, you know.
Jamie Serino (29:44):
So this is going to blow up everything as we know
, completely blow up everythingas we know.
Despina Kontos (29:46):
I wish I was around to see it.
Jamie Serino (29:47):
I don't think we'll be around to see it, but I
wish I mean it definitely isdiscussed a lot, but I guess
separating the hype or the sortof future outlook of it versus
the reality of it.
But I know a lot of companiesare preparing for it and working
on it and they keep saying it'sgoing to happen soon.
But then to your point, howlong would it then become more
(30:07):
mainstream?
Et cetera, et cetera.
But then I just read an articlewhere there would be a concern
even about the electrical grid.
Despina Kontos (30:15):
It's going to be the same.
Jamie Serino (30:16):
Do we have the electrical grid to actually
support all that?
Despina Kontos (30:21):
Well, things will change.
Once upon a time, a computerwas an entire room and now it's
in your pocket, right, so wewill evolve.
You know it's going to bedisruptive, everything you know.
It's just.
Jamie Serino (30:33):
That's the nature of human evolution you know, not
the first time, not the lasttime that something like that
happened yeah, and and do yousee, like even some of the
medical technology making itsway into the home?
Like, I definitely see the consand the danger of this, but
could someone you know givethemselves an X-ray or give
themselves some sort of scan?
(30:54):
Right, it's a part of maybe ahousehold piece of equipment or
something.
Despina Kontos (30:59):
I think, so I'm all for it.
You know, don't tell me there'sno danger of being in the
hospital.
You go with one type of thingand you come out with five
others, right.
So there are dangers in both.
Wherever there's human natureinvolved, there are risks,
because, in my view, we're theones who always mess up
everything.
But yeah, I think all thesethings will happen, absolutely.
(31:24):
I'm excited for it.
Jamie Serino (31:27):
Yeah, I mean cause that's really.
It's like democratization.
You see, this access to thingsand you know, down to the
individual level, um is justwhat has been happening over and
over and over again.
Um, so you, you brought up umscanning and trying to predict.
So we talked a little bit aboutusing the AI to sort of draw
(31:49):
all this information together,doing the laundry for the doctor
, you know, as you you weresaying how about, like, scanning
and using, you know, predictivemetrics and doing some sort of
preventative treatment?
Like I do not have a tumortoday, but chances are because
of my family history.
Maybe there are genetic testsalso and then maybe now ai is
(32:12):
picking up something because ofmy age or something they have an
inflammation in my body orwhatever it is.
Um, and then all of a sudden,red flag, you, you might develop
a tumor somewhere.
Like, are you also doing thatsort of predictive?
Despina Kontos (32:26):
Yes, absolutely Absolutely, and I think it's
important to have thisdiscussion, you know, because
through interventions, lifestyleinterventions, you know, I'm
not saying we're going to liketotally prevent cancer or
whatever, but you know theimpact we could have on
healthcare by lifestyleinterventions.
Battling obesity, you know.
You know wellbeing, you know,and so forth, hypertension, you
(32:48):
know all of these factors thatare water, based on lifestyle
and some genetics.
I'm not like saying just alllike lifestyle, but you know, we
can have a you know impact onpublic health.
That is, could have more impactthan you know, for example, all
cancer therapists combined right.
So, and so I think it'simportant to have these
(33:11):
discussions around preventionand and and, for people to be
open to the idea of if, if theywant that in their lives, to
have that information.
Like you know I have a lot ofdiscussions with friends or
colleagues.
I'm like you know, I'm theperson who likes to know, right?
(33:31):
You know I want to do genetictesting.
I want to know what's going onand they're like why do you want
to know if you're going to have, like, alzheimer's?
Well, I want to know because Iwant to be on the lookout.
For example, right Like there'sa whole bunch of therapies now
in development.
You know preventative things.
You know lifestyleinterventions.
If nothing else, I want toprepare myself.
(33:52):
I think it's important to havethat information and have these
conversations.
Peter Carucci (34:01):
You know, as you're saying, that I'm just
thinking about a family member Ihad, who he's since passed, but
he had a certain condition andhe got a scan back and the
doctor said to him hey, I've gotgood news.
This measurement did not go downanymore in the last month, and
(34:22):
so he celebrated by going out tohave the biggest dinner and
eating everything he shouldn'thave eaten in his celebratory
meal.
And I said to him what are youdoing?
The doctor said this isn't inyour diet.
This isn't in your diet.
You're not supposed to eat this.
You got to have more broccoli.
What are you doing?
And he goes hey, I got greatnews.
The prediction came back that Ihaven't lost any more of that
(34:46):
thing.
And this is great and it'sfunny.
You can lead the horse to waterwith a lot of this information,
but you can't make that horsemake the water.
So with humanity, it'sinteresting that AI can make
these predictions and then wemay just go in and mess it right
up.
We might just go in and youknow.
(35:07):
So do you find that?
I mean, do you see that battleplaying out?
Despina Kontos (35:13):
I see it I don't know if I see it as a battle.
You know I'm a huge believer inpersonal choice and personal
responsibility and personalresponsibility.
You know there's a very goodargument of you know the
extending life okay, as aconcept.
(35:33):
You know, I want you to preventthese things and that because I
want you to live longer.
Okay, how much value does thathave to the individual?
Maybe, if some person wants tolive a certain way, it'll be
shorter you know, I respect that, you know or they want to eat
this all day.
They eat all this all day.
Like, what can I tell you?
I'm just going to give you theinformation.
You do what you feel isappropriate for you to do.
I'm not going to tell you, I'mnot going to force you to do
(35:55):
anything.
I'm very against like forcingin that sense.
But I do think it's importantto have the information and
aware, well-informed choices andthen if your strategy is an
exit strategy, that's on youyeah, I, I think you know
(36:15):
there's it, it is, it is on youand, and, I think, different
people.
Jamie Serino (36:21):
I've heard people say, like you know, the mental
health community, that there'sactually an overwhelming amount
of information now about ourhealth.
And you have these sort ofbiohackers, influencers that you
know spend almost every minuteof their day doing something
that is extending their life,you know, or fighting against
your life, you know, and that'sthat's great.
(36:42):
That's one extreme.
But then you know, some peoplethat you know maybe have anxiety
issues are beginning.
You know, am I drinking enoughwater?
Am I exercising enough?
Am I getting enough hours ofsleep?
Am I meditating?
Is my cortisol levels too high?
Are my cortisol levels too high?
And so, yeah, there can be anoverwhelming amount of
information, but I guess then atthat point, each person has to
(37:06):
deal with it and pick and choosewhat they want to incorporate
into their lives.
You know, I guess to that thisis making me think to ask the
question of, like you know, doesthis add, you know, an even
more overwhelming amount ofinformation to the person?
It's just like, hey, theinformation is there, each
individual deals with it, youknow, is that what it is, or is
(37:27):
there some other you know thinghere, anything else that you'd
have to add?
It is like or is there someother you know thing here,
anything else that you'd have toadd?
To some, to something like likethat, that type of future I?
Despina Kontos (37:37):
can only speak for myself and my own experience
.
Right, and at the end of theday, I'm not even a doctor, I'm
just a computer scientist and anengineer.
I personally like theinformation.
I don't feel overwhelmed by theinformation, but I think you
know one of the things we don'tdo very well in our society is
(37:58):
to teach people in oureducational system whatever it's
very like how to use theirbrain, right, we are very kind
of like information focused,like you have to learn, learn,
learn things, but we don'treally teach people how to use
their brain.
In what sense?
Like you know, a lot of whatyou're talking, just talked
(38:19):
about, is like how you sufferfrom your own brain, like how
you can't even untangle your ownthoughts, how you let your
brain drag you down in anydirection, and how we live our
lives in ability and the skillto pause and be the masters of
(38:54):
our own brain, of our ownemotions and, at the end of the
day, in our own destiny.
We are just like being draggedalong, right?
So I think so from myperspective.
So I think so from myperspective.
I like information, I want toknow.
That's always been my thing.
Like I always want to know.
(39:17):
I don't care, I want to know.
I don't care the cost, I wantto know.
But I think it's important to,as a society, to develop the
skills to use our brains and ouremotions in a constructive,
healthy, life-sustaining way.
We don't do that, you know.
(39:37):
Most of what we suffer from isour own thoughts and our own
emotions about something thatdoesn't even exist, like either
something that has happened inthe past or something that we
think that will happen in thefuture.
But we very rarely are presentin the present moment.
Peter Carucci (39:53):
Yeah.
Despina Kontos (39:53):
So, that takes us in a very, completely
different kind of direction thisdiscussion.
But the problem is how?
Is not the information or thetools or whatever is how.
The problem is always us, it'salways it's us.
You know our individualresponse to what is happening.
(40:14):
That is always the problem inour existence, right, I think at
least, and you know our abilityto respond to what is happening
.
That, I think, is where we needto focus as a society and take
control of our ability torespond in a conscious, aware
and health sustaining manner.
That is my view.
Jamie Serino (40:36):
Yeah, it's well said.
I think it involves a littlebit of Taoism, a little bit of
Buddhism.
Yeah, the thought comes in, thethought leaves Right, and you
know it also makes me thinkabout going back to what you
said earlier.
There's a bit of like changemanagement here that needs to
happen as you introduce this newthing, and we talked about
(41:00):
introducing it to the healthsystem, but even introducing it
to doctors, any anyone using,you know, this medical
information, and thenintroducing it even to the
people, the patients and stuff.
So the the change managementpiece is you know how do we use
this information?
Um, either as a practitioner oras the patient receiving it.
(41:22):
Um, and that change managementpiece, you know that that that's
where like disruption comes in,and disruption is successful
when the change management partgoes well.
Um, when it doesn't go well,that's where like disruption
comes in and disruption issuccessful when the change
management part goes well.
When it doesn't go well, that'swhen it gets all kind of clunky
.
So I mean, do you ever havethose types of discussions in
terms of like, okay, now we'reintroducing this new thing, you
(41:43):
know what's going to be theresistance to it.
How do we get over thatresistance?
That's the change managementand organizational, that's the
behavioral Exactly.
Despina Kontos (41:51):
Exactly, we have these discussions all the time
and again, as I said, I'm justthe engineer, guys.
I am making the gadgets andbringing them to the doorstep.
That is my expertise, mytraining, my job, my calling, my
thing.
I make the gadgets and bringthem to your doorstep.
But I think the process ofevaluating whether or not
(42:11):
something should actually beused by a physician should
actually be communicated to thepatients.
That's a whole differentscientific field.
Like there is what we call likeimplementation science,
comparative effectivenessscience, outcomes research.
There are actual experts aboutthis and we kind of bypass that
whole thing.
(42:32):
We bypass that thing.
We just take it from me, theengineer what do I know?
And try to like shove it in theface of the health system with
the hands of the physician.
We bypass the whole process,folks.
Like we need to engage theseexperts to do their job right
and we need to work together.
Like I'll give you the gadget.
You take this gadget, see if itmakes sense or not.
(42:53):
If you have some feedback, giveit back to me.
I'll tweak it, I'll change it,I'll do whatever it takes you
know, and then the health systemneeds to be communications Like
.
there needs to be anunderstanding that, at the end
of the day, we need to do thisto better our society and
healthcare and health patients.
At the end of the day, we needto do this to better our society
and health care and healthpatients at the end of the day,
(43:14):
right.
If this is not our goal, wewill fail as a society in this,
and so we need to have thesekind of open communications and
coordination between people likeme.
Peter Carucci (43:36):
And coordination between people like me, people
who are in the middle and peoplewho are on the other side, to
be able to successfully bringthis to help us, to help us as
humanity right.
And you're currently using, orrather you created, this kind of
system and you're working withAI to make certain predictions
about specifically certaincancers and whatnot.
About specifically certaincancers and whatnot, and I
recall in an earlierconversation we had, you're also
looking at it in terms of like,maybe predicting Alzheimer's.
Despina Kontos (43:55):
Yeah, not me, my colleagues, right, I work in
cancer specifically, butcolleagues of mine.
Peter Carucci (44:01):
I guess this is a field kind of the predictive
Generally, overall, Exactly, I'dlove to hear a little more
about what you see that futurelooking like, or what that looks
like.
Despina Kontos (44:12):
I think there are similar considerations
across the board.
If it's cancer, if it'sAlzheimer's, if it's cardiac
disease, there's similar trendsand similar work happening
across the board.
And we have the data, thetechnology, the expertise to
develop these technologies andthese tools that can help us
(44:32):
personalize care.
And so, again, the questionbecomes what are the barriers,
the opportunities?
And we need to decide as asociety, how do we want to use
these technologies?
I think it's up to us to decidecollectively what we want to do
with it.
Peter Carucci (44:50):
Do you really think it's up to us to decide
collectively what we want to dowith it?
Do you really think it's up tous to decide, or do you think
it's?
Despina Kontos (44:59):
the financial impact.
I know Life is complicated andthere's all kinds of interests,
guys, but again, from myperspective, if we don't all
take personal responsibility, weare doomed.
It's very easy to say it'ssomebody else's fault.
That doesn't help anybody.
It's our collective, individualresponsibility that's
responsible for everything.
We vote, we elect the peoplewho govern us.
(45:22):
We buy products.
We don't buy products.
We invest this whole thing.
I've never been a person wholike, oh, it's like the.
They right, they, they want todo this.
They, who are they?
It's us.
I don't know this.
They.
I've never seen this they.
Have you seen this?
They?
So I think it's.
We all have to take personalresponsibility.
(45:42):
There's no other way out.
Peter Carucci (45:44):
But in this polemic we're discussing the
possibility of, let's say Idon't want to say names but a
large company Amazon or Googleor any large company could just
jump the shark here, take thistechnology and try to force its
hand into healthcare for thefinancial benefits.
(46:05):
Are you the end?
Despina Kontos (46:06):
user of this.
They all depend on us.
They all depend on the end user?
Peter Carucci (46:09):
Who are they depending on?
Who are they going to sell itto?
Despina Kontos (46:11):
It's your money and my money, and you know who's
going to pay for this right,who's going to pay for it.
So we don't diminish the roleof personal responsibility.
It is the only way out.
The only way out, I think Allright?
Jamie Serino (46:29):
Well, Despina, as we look to wrap up here, is
there anything that we didn'task or anything that you would
want to make sure to add?
Despina Kontos (46:39):
You know, I think the future is bright.
If we want to make it bright.
The future can be very doom andgloomy if we want to make it
doom and gloomy.
History has shown that, youknow, us humans are probably the
worst kind on the planet.
But it has also shown that weare probably the worst kind on
the planet, but it has alsoshown that we're probably the
best kind on the planet as well.
So I think it's up to us todecide how we want to flip that
coin.
Yeah, and, of course, everybodyin their daily lives to take
(47:02):
personal responsibility foreverything that they do, as
little as or big, whateverthey're doing, every moment
counts, I think in this game.
Jamie Serino (47:13):
Yeah, and that's a really good message because,
you know, of course we didn'treally touch upon this, but
there's always these likeend-of-the-world scenarios with
AI, terminator, you know,missiles launching.
And it's interesting because,you know, we talked about Asia
and the Asian cultures tend notto have that mythology and the
(47:34):
Western cultures tend to havethat mythology that the robots
are going to rise up and take usover, and you know, and you
know, I don't know who's rightthere, but you're sort of saying
that the individuals have theresponsibility and they have the
wherewithal and they have thecontrol.
Despina Kontos (47:48):
Really, in the end, Show me a time in human
history where robots were nottaking over Quote robots when
there was peace and prosperityand that everybody was in La La
Land.
You know human history is, youknow, continuously in turmoil,
mayhem, wars, this disaster Like.
(48:11):
Show me a time in human historywhere this does not exist.
So I think there's a little bitof a hype there and, at the end
of the day, if the robots takeover and destroy us, we'll
probably deserve it, you know,because we let them do it.
Jamie Serino (48:23):
We were asking for it.
Despina Kontos (48:25):
You know, I also have a perspective.
Why are we always doom andgloom Like the robots are going
to be like the worst of us.
There's a fair chance that thatmay be the best of us, you know
the best?
yeah, I think we're totallycapable of destroying this
planet without any help from therobots.
We don't need the robots toself-destruct.
I think we're doing a good jobat it as it is.
(48:46):
So, who knows, maybe the robotsare going to save us because
they might tap into the best ofour intelligence.
That's another flip side of thestory.
That's a scenario.
Jamie Serino (48:57):
I like that A message of control, control and
Despina.
It's been fascinating.
We learned so much here.
We're probably gonna have tocome back and talk to you some
more because we still have, Ithink, a thousand questions, but
I want to thank you.
I want to thank everyone forwatching and listening and we
will see you all next time.
Despina Kontos (49:17):
Thank you.
Thank you for having me andlet's get together again before
the robots come.
Okay, thank you All right.
Jamie Serino (49:24):
Thank you.
Popular Podcasts
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
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 Clay Travis and Buck Sexton Show
The Clay Travis and Buck Sexton Show. Clay Travis and Buck Sexton tackle the biggest stories in news, politics and current events with intelligence and humor. From the border crisis, to the madness of cancel culture and far-left missteps, Clay and Buck guide listeners through the latest headlines and hot topics with fun and entertaining conversations and opinions.