DOI: h p://doi.o g/10.12795/A gumen os/2024.i27.07
A gumen os de Razón Técnica, n.º 27, 2024, pp. 215-230
TELEMEDICINE AND DIAGNOSTIC ARTIFICIAL INTELLIGENCE:
EVOLUTION, PROSPECTS, AND SOCIAL CHALLENGES
TELEMEDICINA E INTELIGENCIA ARTIFICIAL DIAGNÓSTICA:
EVOLUCIÓN, PERSPECTIVAS Y DESAFÍOS SOCIALES
REGINA PENNER
Sou h U al S a e Uni e si y, Russia
penne . [email p o ec ed]
h ps://o cid.o g/0000-0002-3277-7274
RECIBIDO: 30/09/2024
ACEPTADO: 15/11/2024
Abs ac : This a icle examines he e olu ion o elemedicine as one o he key
ields in mode n medicine, u ilizing in o ma ion and communica ion echnologies
o p o ide medical se ices emo ely. The s udy’s goal is o analyze he his o ical
de elopmen o elemedicine and i s impac on medical p ac ice, as well as assess
he p ospec s o using Diagnos ic A i icial In elligence (Diagnos ic AI) in
elemedicine. The esea ch me hodology includes discou se analysis, based on he
wo ks o Michel Foucaul and Gilles Deleuze, allowing he examina ion o
medical ins i u ions as mechanisms o con olling and no malizing pa ien
beha io . The s udy concludes ha elemedicine signi ican ly imp o es access o
heal hca e, educes cos s, and enhances ea men e iciency. Diagnos ic AI
ep esen s a new phase in elemedicine, capable o accele a ing and imp o ing
diagnos ic p ocesses, hough i s implemen a ion is accompanied by signi ican
e hical and social challenges. The no el y o his a icle lies in c i ically examining
he ole o elemedicine and diagnos ic echnologies in con empo a y socie y, wi h
a ocus on hei disciplina y unc ions and he in luence on he ela ionships
be ween pa ien s and heal hca e p o essionals. The conclusion emphasizes he
impo ance o u he s udy and egula ion o Diagnos ic AI in medical p ac ice o
minimize isks and ensu e equi able access o heal hca e.
Keywo ds: elemedicine; a i icial in elligence; diagnos ic a i icial in elligence;
Google; AMIE; OpenAI; e olu ion; p ospec s; social challenges.
REGINA PENNER
A gumen os de Razón Técnica, n.º 27, 2024, pp. 215-230
216
Resumen: Es e a ículo examina la e olución de la elemedicina como uno de los
campos cla e en la medicina mode na, u ilizando ecnologías de la in o mación y
la comunicación pa a p opo ciona se icios médicos de o ma emo a. El obje i o
del es udio es analiza el desa ollo his ó ico de la elemedicina y su impac o en la
p ác ica médica, así como e alua las pe spec i as de uso de la In eligencia
A i icial Diagnós ica (IA Diagnós ica) en la elemedicina. La me odología de
in es igación incluye el análisis del discu so, basado en los abajos de Michel
Foucaul y Gilles Deleuze, lo que pe mi e examina las ins i uciones médicas
como mecanismos pa a con ola y no maliza el compo amien o del pacien e. El
es udio concluye que la elemedicina mejo a signi ica i amen e el acceso a la
a ención médica, educe los cos os y mejo a la e iciencia del a amien o. La IA
diagnós ica ep esen a una nue a ase en la elemedicina, capaz de acele a y
mejo a los p ocesos de diagnós ico, aunque su implemen ación es á acompañada
de impo an es desa íos é icos y sociales. La no edad de es e a ículo adica en
examina c í icamen e el papel de la elemedicina y las ecnologías de diagnós ico
en la sociedad con empo ánea, con un en oque en sus unciones disciplina ias y la
in luencia en las elaciones en e pacien es y p o esionales de la salud. La
conclusión en a iza la impo ancia de segui es udiando y egulando la IA
diagnós ica en la p ác ica médica pa a minimiza los iesgos y ga an iza un acceso
equi a i o a la a ención médica.
Palab as cla e: elemedicina; in eligencia a i icial; in eligencia a i icial
diagnós ica; Google, AMIE; OpenAI; e olución; pe spec i as; e os sociales.
In oduc ion. The Social Ins i u ion o Medicine
The phenomenon o medicine and he discou se su ounding i a e
no s a ic o comple e p ojec s; hey e ol e o e ime, ac oss a ious
cul u es and socie ies. Fo ins ance, one majo e ela ion o
Wes e n Eu opean socie y a he u n o he 19 h and 20 h cen u ies
was he ecogni ion o emale sexuali y (e.g., in he wo ks o
Sigmund F eud, such as “S udies on Hys e ia” o “Th ee Essays on
he Theo y o Sexuali y”). Michel Foucaul , a guably he mos
enowned “a chaeologis o knowledge,” p oposed he concep o
he clinical gaze in his book “The Bi h o he Clinic” (Foucaul ,
2003). Th ough his lens, we obse e a shi om iewing diseases
TELEMEDICINE AND DIAGNOSTIC ARTIFICIAL INTELLIGENCE
A gumen os de Razón Técnica, nº 27, 2024, pp. 215-230
217
as abs ac en i ies o s udying illness as some hing isible and
diagnosable wi hin he human body.
This ansi ion, on one hand, ans o ms he phenomenon o
medicine i sel and i s social p esence, while on he o he hand, i
se s he di ec ion o he discu si e p ac ices and social ela ions ha
o m a ound i . Following Foucaul ’s insigh s, we can ace se e al
s ages in he his o ical ans o ma ion o medical discou ses:
1) Classical Medicine (17 h – 18 h cen u ies): Cha ac e ized by a
pu ely medical iew o disease, whe e diseases we e ea ed as
indi idual en i ies, independen o he pa ien . Physicians, inspi ed
by na u al sciences, aimed o classi y and sys ema ize diseases,
much like bo anis s classi y plan s. Symp oms we e conside ed
ex e nal mani es a ions o in e nal diseases, and he physician’s ask
was o in e p e hese symp oms accu a ely.
2) Clinical Medicine (La e 18 h – Ea ly 19 h cen u ies): Du ing
his pe iod, he clinic as a disciplina y ins i u ion was bo n. The ocus
shi ed om abs ac diseases o he speci ic body o he pa ien ,
whe e diseases could be localized, connec ed o ce ain o gans and
issues, and diagnosed, e en h ough dissec ion. This new pa adigm
o bodily obse a ion con as ed wi h he pa ien ’s subjec i e
expe ience.
3) Biopoli ics (19 h – 20 h cen u ies): Medicine became a ool o
powe , aimed a con olling indi idual bodies and popula ions. A
his s age, s a e heal hca e ins i u ions we e es ablished, se ing no
only o ea pa ien s bu also as mechanisms o con olling public
heal h (Foucaul , 2008). Fo example, psychia y o ms he
unde s anding op ics h ough which he no ms o beha io and he
deg ees o de ia ion om hem a e illumina ed; pa ien s o
psychia ic clinics in such op ics a e ans o med in o objec s o
disciplina y con ol, whe e he ask o medicine is no so much in
ea men as in he no maliza ion o de ian beha io . In sum,
medicine is in eg a ed in o he s a e machine, o ien ed owa d
REGINA PENNER
A gumen os de Razón Técnica, n.º 27, 2024, pp. 215-230
218
main aining and e en imp o ing he heal h o he na ion, bu mainly
in s a is ical indica o s.
La e , poin ing o he ansi ion om disciplina y sys ems o
socie ies o con ol, G. Deleuze de eloped he Foucauldian concep ,
no ing he medicaliza ion o e e yday li e, when e e yday p ac ices
(nu i ion, physical ac i i y, sleep) become objec s o medical
a en ion, mo ing om he p i a e o he public sphe e, mo e
accu a ely subjec o measu emen and con ol; and he
echnologiza ion o medicine, as a esul o which medicine is
ans o med in o a complex ne wo k, whe e heal h managemen is
ca ied ou h ough in o ma ion pla o ms ha collec and analyze
da a abou pa ien s in eal ime (Deleuze, 1992).
This echnological u n is he implemen a ion o elemedicine,
which is now a legi ima e a ea o medicine ha uses con empo a y
in o ma ion and communica ion echnologies o p o ide medical
se ices and pa ien consul a ions a a dis ance. Telemedicine has
main ained i s ele ance o e he pas decade; he mos s iking
illus a ion o his is s ill he COVID-19 pandemic, when access o
adi ional heal h se ices became di icul a ound he wo ld.
The so-called elemedicine “body” includes indi ec
consul a ions, pa ien heal h moni o ing, eme gency cases and
educa ional p og ams as in eg al elemen s. Fo example,
consul a ions allow pa ien s o ecei e answe s o hei que ies om
doc o s online, wi hou he need o di ec physical con ac . The
impo ance o such a p ocedu e inc eases o people li ing in emo e
o ha d- o- each a eas. Bu consul a ions also help educe he ac ual
load on medical ins i u ions (sup a-indi idual u ili y) and educe he
wai ing ime o an appoin men (indi idual u ili y).
Dissemina ion o sma de ices ac oss indi idual households is
ano he elemen o elemedicine. Using mobile de ices and senso s,
pa ien s can emo ely moni o key heal h indica o s (blood p essu e,
blood suga , hea a e). This da a is au oma ically ansmi ed o
TELEMEDICINE AND DIAGNOSTIC ARTIFICIAL INTELLIGENCE
A gumen os de Razón Técnica, nº 27, 2024, pp. 215-230
219
doc o s, which allows hem o p omp ly espond o de ia ions om
he no m. In eme gency si ua ions, elemedicine echnologies allow
he ic im o p omp ly con ac a doc o who can p o ide i s aid,
gi e ecommenda ions o s abilizing he condi ion o decide on he
need o hospi aliza ion. Mo eo e , oday elemedicine is
inc easingly used o ain medical wo ke s. Online cou ses, webina s
and o he educa ional esou ces help imp o e he quali ica ions o
medical pe sonnel and in o m he popula ion.
Telemedicine is a spec acula illus a ion o e ol ing disciplina y
ins i u ions; i signi ican ly imp o es access o medical ca e o he
gene al popula ion, educes heal h ca e cos s, and inc eases he
e ec i eness o ea men . A he same ime, he need o ensu e he
secu i y o pe sonal da a, p o ec pa ien p i acy, and s anda dize
elemedicine se ices is g owing exponen ially. In addi ion, o e he
pas i e yea s, non-human ac o s ha e inc easingly been included
in he doc o -pa ien ela ionship. As ecen ly as he end o he 20 h
cen u y, he unc ions o a machine in a medical con ex we e limi ed
o analyzing he da a i ecei ed. Today, a machine claims o be a
“specialis ” capable o p o iding p o essional se ices in consul ing
a pa ien , diagnosing him, and suppo ing his ea men .
Diagnos ic AI: A New Miles one in Telemedicine Opens New
Challenges
A i icial in elligence implemen ing diagnos ic goals in medical
discou se (so-called Diagnos ic AI) is one o such agen s. Diagnos ic
AI is one o he mos p omising a eas in medicine and in o ma ion
echnology. I s applica ion co e s a ange o asks: om in e p e ing
images and analyzing gene ic da a o p edic ing ea men esul s and
suppo ing medical decisions.

REGINA PENNER
A gumen os de Razón Técnica, n.º 27, 2024, pp. 215-230
220
A i icial in elligence in medicine began wi h he i s expe
sys ems in he 1960s. Howe e , a echnological leap has only been
made in he las decade in he con ex o machine lea ning, in he
a eas o deep lea ning and big da a p ocessing. Mode n AI-based
diagnos ic sys ems a e able o analyze huge amoun s o da a, which
signi ican ly exceeds human capabili ies, and p o ide doc o s wi h
ecommenda ions based on his da a. One o he key b eak h oughs
has been he de elopmen o AI sys ems o analyzing medical
images. Fo example, a s udy by analys s om India showed ha
such sys ems can iden i y pa hological changes in images wi h an
accu acy compa able o o e en supe io o he capabili ies o
expe ienced adiologis s ( he a icle speci ies ha sys ems ained on
MRI o CT da a a e able o ecognize cance in he ea ly s ages,
which signi ican ly inc eases he chances o success ul ea men )
(Ma hu e al., 2022).
The use o AI in diagnos ics has a numbe o ob ious ad an ages.
Fi s ly, i is he high speed and accu acy o da a p ocessing, which is
especially impo an in he con ex o a sho age o quali ied medical
pe sonnel and an inc easing bu den on heal hca e (Pin o-Coelho,
2023). Secondly, AI can help minimize he human ac o , which
o en leads o diagnos ic e o s (Russell & No ig, 2021). Howe e ,
he implemen a ion o Diagnos ic AI is associa ed wi h a numbe o
challenges and limi a ions. One o he main p oblems is he us o
doc o s and pa ien s in AI sys ems. Ex ensi e clinical ials and
ce i ica ion o such sys ems a e needed o ensu e hei eliabili y and
sa e y. This o malizes he e hical p oblems in he use o AI in eal
medical p ac ices, ounded, on he one hand, by he con iden iali y
o pa ien da a, and on he o he hand, by he ques ion o he
esponsibili y o AI o he decisions made (in gene al, abou he
legi imacy o applying he c i e ia o esponsibili y o so wa e)
(Flo idi, 2023).
TELEMEDICINE AND DIAGNOSTIC ARTIFICIAL INTELLIGENCE
A gumen os de Razón Técnica, nº 27, 2024, pp. 215-230
221
In any case, he e hical issues su ounding he use o diagnos ic
AI ouch on se e al a eas a once. One o he key issues is ensu ing
ai access o new echnologies. The e is a isk ha such sys ems will
be used p edominan ly in de eloped coun ies, while de eloping
coun ies may be le wi hou access o hese achie emen s ( he e y
o mula ion o he p oblem epea s he classic plo o class s uggle,
ansc ibed by McKenzie Wa k on he “hacke ” ma e ial (Wa k,
2004). An equally undamen al ques ion is who is esponsible o he
decisions made by AI. I he sys em makes a mis ake, who will be
held accoun able: he de elope s, he doc o s using he sys em, o
he in elligen sys em i sel (Jobin, e al., 2019)?
AMIE: P elimina y esul s o es ing he Google’s neu al
ne wo k
Diagnos ic AI is a e olu iona y ool ha has he po en ial o
signi ican ly imp o e he quali y o medical ca e and make
diagnos ics mo e accessible and accu a e. Howe e , success ully
implemen ing AI in medical p ac ice equi es add essing a numbe
o echnical, e hical, and social issues. A icula e Medical
In elligence Explo e (AMIE), Google’s medical neu al ne wo k,
demons a es his pa adox o echnological de elopmen and e hical
u bulence.
The de elopmen and es ing o AMIE, an inno a i e sys em
based on la ge language models (LLM) designed o conduc medical
diagnos ic dialogues, is discussed in he pape “Towa ds
Con e sa ional Diagnos ic AI” by esea che s om Google
Resea ch and Google DeepMind (Tu, e al., 2024). The au ho s
de ine he goal o he s udy as he c ea ion o a i icial in elligence
capable o imp o ing he a ailabili y and quali y o medical ca e by
in oducing ad anced echnologies in o he diagnos ic p ocess and
REGINA PENNER
A gumen os de Razón Técnica, n.º 27, 2024, pp. 215-230
222
gene a ing AI ecommenda ions o he ea men o a speci ic
pa ien (Tu, e al., 2024, p. 3). They cla i y ha dialogue be ween a
doc o and a pa ien is he basis o medical p ac ice. Mo eo e , up o
80% o diagnoses a e made on he basis o anamnesis ob ained
du ing communica ion wi h he pa ien (see, o example, Pe e son,
1992). Howe e , acco ding o expe s, he con empo a y heal hca e
sys em a ound he wo ld aces many challenges, including a sho age
o quali ied doc o s, especially in emo e and disad an aged a eas,
as well as he high cos o medical se ices (see, o example, Roshan
& Rao, 2000). In esponse o he iden i ied p oblems, he AMIE
neu al ne wo k was de eloped, which uses a mode n a chi ec u e on
LLM base. A he inpu AMIE ecei es ex in o ma ion in he o m
o a ques ion o complain o he pa ien ; based on he ini ial
in o ma ion and answe s o cla i ying ques ions a he ou pu he
neu al ne wo k issues om wo diagnoses co esponding o he
indica ed symp oms, cla i ies he mos p obable o hem and o e s
he pa ien speci ic p esc ip ions.
AMIE was ained using a ich da ase . These included eal
medical dialogues o c ea e a da abase ha closely esembles eal-
li e doc o -pa ien in e ac ions; mul iple-choice medical ques ions o
imp o e AMIE’s abili y o gene alize and sol e clinical p oblems o
a ying complexi y; and expe medical summa ies o p o ide a
sys em wi h eliable and alida ed in o ma ion o diagnosis and
decision-making (Tu, e al., 2024, p. 4–5).
One o he key aspec s o AMIE aining was he sel -lea ning
me hod, which is widely used in mode n neu al ne wo ks (see, o
example, OpenAI, e al., 2024 o Google, 2024). In a simula ed
en i onmen , AMIE in e ac s wi h o he AI agen s, simula ing
in e ac ions wi h pa ien s and ecei ing cons an eedback. This
allowed he sys em no only o eco d e o s, bu also o imp o e i s
skills wi h each new dialogue, g adually mo ing owa ds he expe
le el. To es he e ec i eness and accu acy o he AMIE sys em, a
TELEMEDICINE AND DIAGNOSTIC ARTIFICIAL INTELLIGENCE
A gumen os de Razón Técnica, nº 27, 2024, pp. 215-230
223
blinded s udy was conduc ed in which he sys em was compa ed
wi h p ima y ca e physicians based on in e ac ions wi h “simula ed”
pa ien s. The esul s showed ha AMIE demons a ed highe
diagnos ic accu acy compa ed o doc o s di ec ly. In pa icula , he
sys em ou pe o med physicians on 28 o 32 pe o mance indica o s
assessed by expe s and on 24 o 26 indica o s assessed by simula o s
playing he ole o pa ien s (Tu, e al., 2024, p. 12–15). These esul s
con i m ha AMIE is capable o no only suppo ing high-quali y
medical dialogues, bu also making a signi ican con ibu ion o
imp o ing he quali y o diagnos ics, especially in cases whe e he
physician may be limi ed in ime o esou ces.
The au ho s o he s udy emphasize he ad an ages o using
AMIE, including:
1) A ailabili y: AMIE can ope a e a ound he clock, p o iding
suppo o pa ien s ega dless o ime and loca ion; 2.
2) Da a consis ency and objec i i y: he sys em minimizes human
e o s associa ed wi h a igue, s ess, o pe sonal biases o medical
pe sonnel;
3) Po en ial o lea ning and sel -imp o emen o he neu al
ne wo k i sel : AMIE is able o adap o new da a and medical
disco e ies, which makes i a p oduc i e ool in he con ex o
cons an ly e ol ing medicine (Tu, e al., 2024, p. 17–21).
A he same ime, he esea che s no ed se e al limi a ions
iden i ied du ing he ope a ion o he neu al ne wo k. The key one
was he ex in e ace on which AMIE was launched; messages abou
he pa ien ’s condi ion in ex o ma we e a ed by doc o s as
unusual and incon enien ; adi ional me hods o in e ac ion wi h
pa ien s, di ec con e sa ion, emain amilia o hem. Despi e he
success ul esul s in he simula ion, he au ho s poin ou he need
o u he es ing o he sys em in eal condi ions o assess i s
p ac ical applicabili y and sa e y. And in his case, as no ed ea lie ,
REGINA PENNER
A gumen os de Razón Técnica, n.º 27, 2024, pp. 215-230
230
Google. (2024). PaLM 2. Technical
Repo . h ps://ai.google/s a ic/documen s/palm2 ech epo .pd
Jobin, A. e al. (2019). The Global Landscape o AI E hics
Guidelines. Na u e Machine In elligence, 1(9), 389–399.
h ps://doi.o g/10.1038/s42256-019-0088-2
Ma hu , P., Mish a, S., Awas hi, R., Cywinski, J., e al. (2022).
A i icial In elligence in Heal hca e: 2021 Yea in Re iew.
h ps://doi.o g/10.13140/RG.2.2.25350.24645/1
OpenAI, e al. (4 Ma . 2024) GPT-4 Technical Repo .
h p://a xi .o g/abs/2303.08774
Pe e son, M. C., Holb ook, J. H., Von Hales, D., Smi h, N. & S ake ,
L. (1992). Con ibu ions o he his o y, physical examina ion, and
labo a o y in es iga ion in making medical diagnoses. Wes e n
Jou nal o Medicine, 156, 163.
Pin o-Coelho, L. (2023). How A i icial In elligence Is Shaping
Medical Imaging Technology: A Su ey o Inno a ions and
Applica ions. Bioenginee ing, 10(12), 1435.
h ps://doi.o g/10.3390/bioenginee ing10121435
PRIME S udy P og ess Upda e (2024). Second Pa icipan .
h ps://neu alink.com/blog/p ime-s udy-p og ess-upda e-second-
pa icipan /
Roshan, M., & Rao, A. (2000). A s udy on ela i e con ibu ions o
he his o y, physical examina ion and in es iga ions in making
medical diagnosis. The Jou nal o he Associa ion o Physicians o
India, 48, 771–775.
Russell, S., & No ig, P. (2021). A i icial In elligence: A Mode n
App oach. Pea son.
S auman, E., & Goodie , B. (2011). The Doc o (s) in House: An
Analysis o he E olu ion o he Tele ision Doc o -He o. Jou nal o
Medical Humani ies, 32(1), 31–46. h ps://doi.o g/10.1007/s10912-
010-9124-2.

TELEMEDICINE AND DIAGNOSTIC ARTIFICIAL INTELLIGENCE
A gumen os de Razón Técnica, nº 27, 2024, pp. 215-230
231
Tu, T., Palepu, A., Schaeke mann, M., Saab, K., e al. (10 Jan. 2024).
Towa ds Con e sa ional Diagnos ic AI.
h p://a xi .o g/abs/2401.05654
Wa k, Mc. (2004). A Hacke Mani es o. Ha a d Uni e si y P ess.