Iuga, Iulia C is ina; Socol, Adela
A icle
Go e nmen A i icial In elligence eadiness and b ain
d ain: In luencing ac o s and spa ial e ec s in he
Eu opean Union membe s a es
Jou nal o Business Economics and Managemen (JBEM)
P o ided in Coope a ion wi h:
Vilnius Gediminas Technical Uni e si y (VILNIUS TECH)
Sugges ed Ci a ion: Iuga, Iulia C is ina; Socol, Adela (2024) : Go e nmen A i icial In elligence
eadiness and b ain d ain: In luencing ac o s and spa ial e ec s in he Eu opean Union membe
s a es, Jou nal o Business Economics and Managemen (JBEM), ISSN 2029-4433, Vilnius Gediminas
Technical Uni e si y, Vilnius, Vol. 25, Iss. 2, pp. 268-296,
h ps://doi.o g/10.3846/jbem.2024.21136
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/317679
S anda d-Nu zungsbedingungen:
Die Dokumen e au EconS o dü en zu eigenen wissenscha lichen
Zwecken und zum P i a geb auch gespeiche und kopie we den.
Sie dü en die Dokumen e nich ü ö en liche ode komme zielle
Zwecke e iel äl igen, ö en lich auss ellen, ö en lich zugänglich
machen, e eiben ode ande wei ig nu zen.
So e n die Ve asse die Dokumen e un e Open-Con en -Lizenzen
(insbesonde e CC-Lizenzen) zu Ve ügung ges ell haben soll en,
gel en abweichend on diesen Nu zungsbedingungen die in de do
genann en Lizenz gewäh en Nu zungs ech e.
Te ms o use:
Documen s in EconS o may be sa ed and copied o you pe sonal
and schola ly pu poses.
You a e no o copy documen s o public o comme cial pu poses, o
exhibi he documen s publicly, o make hem publicly a ailable on he
in e ne , o o dis ibu e o o he wise use he documen s in public.
I he documen s ha e been made a ailable unde an Open Con en
Licence (especially C ea i e Commons Licences), you may exe cise
u he usage igh s as speci ied in he indica ed licence.
h ps://c ea i ecommons.o g/licenses/by/4.0/
This is an Open Access a icle dis ibu ed unde he e ms o he C ea i e Commons A ibu ion License (h p://c ea i ecommons.o g/
licenses/by/4.0/), which pe mi s un es ic ed use, dis ibu ion, and ep oduc ion in any medium, p o ided he o iginal au ho and sou ce
a e c edi ed.
Copy igh © 2024 The Au ho (s). Published by Vilnius Gediminas Technical Uni e si y
ISSN 1611-1699 / eISSN 2029-4433
2024
Volume 25
Issue 2
Pages 268–296
h ps://doi.o g/10.3846/jbem.2024.21136
GOVERNMENT ARTIFICIAL INTELLIGENCE READINESS AND
BRAIN DRAIN: INFLUENCING FACTORS AND SPATIAL EFFECTS
IN THE EUROPEAN UNION MEMBER STATES
Iulia C is ina IUGA , Adela SOCOL
Depa men o Finance-Accoun ing, Facul y o Economic Sciences, “1 Decemb ie 1918” Uni e si y o Alba Iulia,
Alba Iulia, Romania
A icle His o y: Abs ac . In he swi ly ad ancing ield o A i icial In elligence (AI), a ield whe e
e e y coun y aims o keep pace, signi ican dispa i ies a e obse ed in how di -
e en na ions adop AI. This s udy explo es he deep, ye insu icien ly s udied,
e ec s o AI on socie al, economic, and en i onmen al aspec s. I pa icula ly
examines how b ain d ain in luences go e nmen al AI implemen a ion capabili-
ies, add essing a gap in exis ing li e a u e. The s udy in es iga es he in e play
be ween go e nmen AI implemen a ion and b ain d ain, ac o ing in mac o-
economic condi ions, go e nance quali y, educa ional le els, and R&D e o s.
U ilizing 2022 da a om Eu opean Union coun ies, he esea ch employs in-
s umen al- a iables eg essions (2SLS and LIML) o coun e ac endogenei y and
uses clus e ing me hods o ca ego izing coun ies based on hei go e nmen
AI le els, alongside spa ial analysis o de ec c oss-na ional spillo e s and in e -
ac ions. The indings e eal b ain d ain’s de imen al e ec on go e nmen al AI
p epa edness, highligh clus e ing endencies, and iden i y spa ial in e depend-
encies. This pape unde sco es he need o s a egic policy-making and ins i u-
ional e o ms o bols e go e nmen AI capabili ies. I ad oca es o a pa adigm
shi in go e nmen amewo ks pos -New Public Managemen e a, ailo ed o
he new challenges posed by AI. The esea ch, howe e , is limi ed o a single
yea and egion, wi h cons ain s on da a a ailabili y and indica o b ead h.
■ ecei ed 29 May 2023
■ accep ed 1 Feb ua y 2024
Keywo ds: go e nmen AI eadiness, b ain d ain, go e nmen expendi u e, spa ial e ec s, spillo e e ec s, human
capi al.
JEL Classi ica ion: O15, H52, O38.
Co esponding au ho . E-mail: [email p o ec ed]
JOURNAL o BUSINESS
ECONOMICS & MANAGEMENT
1. In oduc ion
The concep o A i icial In elligence (AI), o en u ilized in specialized s udies and inc easingly
p e alen in ecen language, encompasses echnically complex and compu e -o ien ed ide-
as ha a e di icul o succinc ly exp ess. F om a go e nmen al s andpoin , AI ep esen s a
mul i ace ed blend o echnology, policy, and social impac , necessi a ing a mul idisciplina y
app oach o i s success ul in eg a ion and applica ion.
In ci izen se ices, AI’s applica ions ange om handling inqui ies and p ocessing docu-
men s o di ec ing eques s, aiding ansla ions, and d a ing documen s (Meh , 2017). Th ee
p ima y AI applica ions in go e nmen s and ou : Robo ic and cogni i e au oma ion, enabling
he ealloca ion o human labo o mo e alue-added asks h ough echnologies like Robo ic
Jou nal o Business Economics and Managemen , 2024, 25(2), 268–296 269
P ocess Au oma ion; Cogni i e insigh s gene a ion, imp o ing p edic i e capaci ies; and Co-
gni i e engagemen , add essing ci izen que ies (Egge s e al., 2017).
Go e nmen en i ies a e inc easingly adop ing AI, expec ed o undamen ally ans o m
hei ope a ions, om policy-making o he execu ion o daily public se ices. AI p o ides
ools o p edic i e analy ics, decision-making, and p oblem-sol ing, pa icula ly aluable in
con ex s o signi ican unce ain y. U ilizing AI capabili ies allows public sec o o ganiza ions
o enhance agili y, an icipa e ci izen needs, op imize esou ce alloca ion, and imp o e se ice
deli e y quali y (Mikhaylo e al., 2018).
AI p esen s coun ies wi h a i al oppo uni y o ad ance hei economic landscapes, espe-
cially in public se ices. I enables enhanced, pe sonalized se ice deli e y (Robles & Mallin-
son, 2023). I s powe ul da a analy ics capaci y helps go e nmen s de i e insigh s om la ge
da a se s, suppo ing e idence-based policymaking. AI equips go e nmen s wi h powe ul
ools o au oma ing bu eauc a ic asks, ex ac ing insigh s om as da a, and cus omizing
public se ices o mee indi idual needs, hus symbolizing a beacon o imp o ed go e nance,
in o med policymaking, and heigh ened ci izen engagemen (F ățilă e al., 2023).
The Eu opean Union, a unique assembly o coun ies wi h a ied economic and echno-
logical pa hs, o e s a key s udy a ea o examining go e nmen AI eadiness and b ain d ain
wi hin i s uni ied poli ical and economic s uc u e. Unde s anding each membe s a e’s AI
p epa edness is c ucial, as dispa i ies could signi ican ly impac he EU’s collec i e policy and
s a egic di ec ion.
The EU’s p inciple o ee mo emen (A icle 45, T ea y on he Func ioning o he EU
(Eu opean Union, 2012)) highligh s i s unique b ain d ain issues, wi h high-skilled wo ke s
easily eloca ing be ween s a es, di e ing om global pa e ns. Membe s a es ace simi-
la p oblems like aging popula ions (Walke & Mal by, 2012) and upskilling needs (Jacobs,
2023). Examining AI eadiness in his con ex e eals oppo uni ies o collec i e s a egies
and coope a ion.
Gi en he EU’s acknowledgmen o AI as an ins umen al o ce shaping u u e socio-
economic con ou s (Eu opean Pa liamen , 2023), gauging he AI eadiness o i s go e nmen s
is indispensable. “The Go e nmen AI Readiness Index p o ides an o e all es ima e o how
p epa ed each coun y’s na ional go e nmen is o implemen ing AI in public se ice deli -
e y” (Ox o d Insigh s, 2022).
The EU’s policy ha moniza ion, alongside i s membe na ions’ a ied economic and e-
chnological s ages (Calde a o & Blum elde, 2022), poses challenges and oppo uni ies in
ackling b ain d ain and enhancing AI eadiness. The EU’s Coo dina ed Plan on A i icial
In elligence (Eu opean Commission, 2018) encou aged na ions o de elop o in eg a e AI
s a egies. By 2023, all 27 EU coun ies had AI s a egies (OECD, 2023; Eu opean Commission,
2020; Ca h e al., 2017). S udying he EU o e s a comp ehensi e iew on AI eadiness and
b ain d ain dynamics, yielding ele an insigh s in o AI in eg a ion and labo mobili y ends
in he EU.
P io esea ch e eals new eme ging pa adigms in public adminis a ion (pos -New Public
Managemen Pa adigms) ha e been iden i ied ha ocus on he impo ance o echnological
inno a ion and a i icial in elligence o imp o e he deli e y o se ices o ci izens and he
policies implemen ed by he go e nmen (Ojo e al., 2019). These new pa adigms ocus on
angible bene i s and c ea ing public alue h ough in es men in dis up i e echnologies.
In eg a ing AI in o go e nance in oduces inno a ion bu also challenges. S udies indica e
ha go e nmen s s uggle o keep pace wi h AI ad ancemen s, showing a no iceable gap in
public sec o AI go e nance (Wi z e al., 2020). Challenges o en a ise om skill sho ages
270 I. C. Iuga, A. Socol. Go e nmen A i icial In elligence eadiness and b ain d ain: in luencing ac o s and spa ial e ec s...
(Dwi edi e al., 2021), e iden in employees’ limi ed AI and machine lea ning knowledge
(Ojo e al., 2019) and a lack o local AI expe s (Gup a, 2019). The sca ci y o AI p o essio-
nals (Al-Mushay , 2019) and insu icien ad anced echnology educa ion (Mon oya & Ri as,
2019) add o hese di icul ies. Simul aneously, many coun ies, especially de eloping ones,
ace a signi ican “B ain d ain” – skilled wo ke s mig a ing o de eloped coun ies o be e
oppo uni ies. The F agile S a es Index highligh s he economic and de elopmen al impac s
o such human displacemen (Fund o Peace, 2022).
In he con ex o he Eu opean Union (EU), spa ial in e ac ions and spillo e s be ween
membe s a es a e c ucial o unde s anding phenomena like b ain d ain and AI eadiness.
In he EU, coun ies a e in e dependen ; de elopmen s in one can in luence o he s. This in-
e connec edness, inhe en o he EU’s s uc u e, necessi a es a comp ehensi e iew o hese
dynamics, conside ing he collec i e impac a he han isola ed na ional ends.
The EU’s p inciple o ee mo emen o wo ke s os e s a uni ied labo ma ke , ena-
bling p o essionals o pu sue oppo uni ies ac oss bo de s wi h ewe mig a ion ba ie s. This
mobili y ca ies signi ican consequences: p o essionals ans e hei skills, expe iences, and
insigh s as hey mo e (F ankowska & Pawlik, 2022), encou aging c oss-bo de collabo a ions
in esea ch and inno a ion.
The in e connec edness o he EU means ha membe s a es a e con inually obse ing,
lea ning om, and adap ing policies om hei neighbo s (Al uza a e al., 2019). A success ul
policy o ini ia i e in one coun y can se e as a bluep in o o he s.
To ou knowledge, he e has been no esea ch explo ing wi hin he EU membe s a es
how he na ional con ex in luences go e nmen eadiness o adop AI in neighbo ing s a es.
The aim o he s udy is o analyze whe he go e nmen s’ eadiness o a i icial in el-
ligence AI in he Eu opean Union coun ies depends on “b ain d ain”, using Go e nmen
A i icial In elligence Readiness Index as dependen a iable and he Human Fligh and b ain
d ain as he majo independen p oxy. Also, gi en he speci ics o he EU labou ma ke and
he eedom o mo emen o wo ke s, we we e in e es ed in disco e ing possible spa ial
in e ac ions and spillo e e ec s be ween coun ies.
The objec i e o he pape is h ee old:
1. Examines he in luence o b ain d ain on go e nmen AI eadiness in EU coun ies;
2. S udies he exis ence o spa ial e ec s and spillo e e ec s be ween neighbo ing EU
coun ies;
3. Highligh he policies ha can be adop ed o educe he go e nmen AI disc epancy
be ween he EU membe s a es.
This s udy add esses he ollowing esea ch ques ions:
■RQ1: Does he Eu opean Union expe ience a nega i e e ec o b ain d ain on AI go -
e nmen eadiness?
■RQ2: A e he e posi i e e ec s o EU- unded p ojec s o R&D, economic eedom, and
go e nmen spending on AI go e nmen eadiness in he Eu opean Union coun ies?
■RQ3: Does he b ain d ain (along wi h con ol a iables) in a coun y in luences he
go e nmen ’s eadiness o adop AI in he neighbo ing s a es?
While he e is indi ec li e a u e explo ing he indi idual phenomena o b ain d ain and
go e nmen AI eadiness, he e emains a pauci y o esea ch speci ically examining he in-
e play be ween hese wo dimensions, especially wi hin he unique con ex o he Eu opean
Union. The exis ing body o wo k o en ea s b ain d ain and AI eadiness as dis inc en i-
ies, wi h limi ed explo a ion o how hey migh in luence each o he . Mo eo e , he spa ial
in e ac ions and spillo e e ec s wi hin he EU, shaped by i s singula policies such as he
Jou nal o Business Economics and Managemen , 2024, 25(2), 268–296 271
eedom o mo emen , p esen a nuanced backd op ha hasn’ been adequa ely add essed
in cu en s udies. This lacuna in unde s anding becomes pa icula ly signi ican gi en he
apid echnological ad ancemen s and shi ing labo dynamics wi hin he EU. As such, he e
is a p essing need o an in eg a i e analysis ha holis ically examines how b ain d ain migh
impac AI eadiness ac oss go e nmen s and ice e sa, and how hese dynamics mani es in
he EU’s in e connec ed landscape. This s udy endea o s o b idge his gap, o e ing insigh s
ha can in o m bo h policy-making and u u e academic pu sui s.
The p esen s udy b idges gaps in he p e ailing li e a u e on AI in go e nmen and he
phenomenon o b ain d ain. His o ically, AI’s ole in public sec o augmen a ion has been
discussed in isola ion, wi h emphasis on i s po en ial o s eamline ope a ions and ele a e
ci izen engagemen . This s udy illumina es he in ica e in e play be ween b ain d ain and
go e nmen AI eadiness, a dimension unde explo ed in ea lie wo ks. Echoing indings om
Ox o d Insigh s (2019) we unde sco e he c i icali y o human capi al in he AI ealm. This
aligns wi h Docquie and Rapopo (2012), who highligh ed he epe cussions o high-skilled
labo emig a ion on a coun y’s echnical p owess. Ou esea ch de ia es om me ely ac-
knowledging b ain d ain as a challenge, ex ending in o i s di ec and indi ec impac s on AI
adop ion, especially in he Eu opean con ex . Fu he mo e, while pas li e a u e like B ynjol s-
son and McA ee (2014) ouched upon po en ial delays in AI de elopmen due o expe ise
sho age, ou indings del e deepe in o he cascading e ec s o such delays on public se ice
quali y and na ional secu i y. By jux aposing economic eedom, EU- unded R&D p ojec s,
and go e nmen eadiness o AI, we o e a comp ehensi e na a i e ha no only augmen s
exis ing li e a u e bu also cha s ou po en ial ajec o ies o u u e esea ch in he domain.
The inno a ion o his pape is ha we inco po a e i e majo g oups o ac o s (human,
mac oeconomic, go e nance, educa ion and esea ch & de elopmen ) in o he model and
s udy bo h panel e ec s o Eu opean Union s a es, as well as spa ial e ec s and da a clus-
e ing.
The esea ch’s o iginali y lies in i s unique combina ion o geog aphical ocus, me hod-
ological obus ness, emphasis on spa ial in e ac ions, explo a ion o he b ain d ain phenom-
enon, and i s b idge be ween heo e ical indings and policy implica ions.
The emainde o his pape is o ganized as ollows: Sec ion 2 ou lines he cu en s a e
o knowledge in he ield and he esea ch hypo hesis; Sec ion 3 de ails he da a and me h-
odology; Sec ion 4 ocuses on he esul s and discussion; and he pape concludes wi h he
conclusions.
2. Li e a u e e iew
2.1. In e ac ion be ween go e nmen AI and b ain d ain
Go e nmen AI signi ies he use o AI echnologies in he public sec o , enhancing e iciency,
e ec i eness, and decision-making. I possesses he po en ial o ans o m go e nmen ope -
a ions by imp o ing ci izen se ices, educing cos s, and op imizing esou ce use. AI in ol es
compu a ional echniques ha enable machines o lea n, eason, and sol e p oblems, akin
o human cogni ion. Wo ldwide, go e nmen s a e inc easingly in e es ed in embedding AI
in o hei ope a ional ab ic.
AI’s applica ion in go e nmen spans a ious domains: in Public Se ices, i could au o-
ma e asks like applica ion p ocessing and inqui y handling; in Policy and Decision-making,
i could analyze ex ensi e da ase s o policy shaping; in Law En o cemen and Public Sa e y,
272 I. C. Iuga, A. Socol. Go e nmen A i icial In elligence eadiness and b ain d ain: in luencing ac o s and spa ial e ec s...
echnologies like acial ecogni ion and p edic i e analy ics could augmen sa e y e o s; in
In as uc u e and Resou ce Managemen , AI could assis in esou ce managemen and in-
as uc u e planning (Wang & Cui, 2022); and in Heal hca e, AI could imp o e public heal h
ou comes by analyzing medical da a o disease de ec ion and ea men op imiza ion (Go-
mes de Sousa e al., 2019).
Coun ies’ eadiness o AI echnology a ies, as e ealed by an Ox o d Insigh s (2019)
e iew, which iden i ied challenges in adop ing AI o he common good, including policy,
capaci y, and esou ces, wi h human capi al as a key esou ce. The Human ligh and b ain
d ain indica o ocuses on he economic e ec s o skilled labo mig a ion and i s impac on
a na ion’s ad ancemen . B ain d ain, a complex phenomenon, pa icula ly a ec s de eloping
coun ies, wi h skilled wo ke s emig a ing o mo e de eloped egions (Docquie & Rapopo ,
2012). This mig a ion impac s a na ion’s capaci y o main ain and g ow i s expe ise in c i ical
a eas like AI. In he con ex o globaliza ion and in e na ional compe i ion, b ain d ain a ec s
a coun y’s abili y o main ain and de elop i s expe ise in key a eas such as AI and a ac
o eign di ec in es men (Czaika & de Haas, 2015; Sia , 2013).
In he Eu opean Union, he e’s a no able mig a ion o expe s om Cen al and Sou h-Eas-
e n Eu ope o Wes e n coun ies (Bălan & Ol eanu, 2017). This mo emen , d i en by pe sonal
and socio-economic ac o s, began a e he all o communis egimes in he 1990s. The
impac o b ain d ain on go e nmen AI is mul i-dimensional: I es ic s access o needed
expe ise o AI de elopmen (B ynjol sson & McA ee, 2014), po en ially causing delays in AI
adop ion and isks o da a p o ec ion (Dignum, 2019). I also leads o a loss o AI alen o
he p i a e sec o , slowing public sec o inno a ion. Addi ionally, go e nmen s migh become
o e - elian on p i a e AI solu ions (Gesk & Leye , 2022), posing na ional secu i y isks and
loss o con ol o e da a and algo i hms (O e , 2022). The public-p i a e pay gap (Ag awal
e al., 2019) and he lack o public sec o in es men in AI R&D (Johnson, 1965) u he uel
b ain d ain.
O e all, while AI holds ans o ma i e po en ial o go e nmen ope a ions, i s e ec i e
in eg a ion aces hu dles due o b ain d ain, esou ce limi a ions, and e ol ing AI de elop-
men and policy landscapes. This in e play equi es a mul i ace ed app oach o enhance AI
eadiness in he public sec o , conside ing he dynamics o economic dispa i ies and global
alen mobili y.
Conside ing hese, we can o mula e he main esea ch hypo hesis o his s udy:
H1: The e is a nega i e e ec o b ain d ain on AI go e nmen eadiness in he Eu opean
Union coun ies.
2.2. In e ac ion be ween mac oeconomic, go e nance, educa ion and
esea ch & de elopmen a iables on AI go e nmen eadiness
B ain d ain, go e nmen spending, economic eedom and EU- unded esea ch and de elop-
men p ojec s o a i icial in elligence a e undamen al hemes o he con empo a y deba e
on sus ainable de elopmen and na ional o Eu opean compe i i eness.
The in oduc ion o a i icial in elligence in o go e nmen ope a ions has he po en ial
o ans o m he mode he public sec o ope a es, op imizing e iciency, e ec i eness and
decision-making. Go e nmen spending can ha e a posi i e impac on go e nmen eadiness
o AI, ensu ing ha public ins i u ions can eap he bene i s o his inno a i e echnology
(B ed , 2019).
Jou nal o Business Economics and Managemen , 2024, 25(2), 268–296 273
In es ing in in as uc u e, heal h, educa ion and esea ch can help inc ease a coun y’s com-
pe i i eness and imp o e ci izens’ quali y o li e (Bose e al., 2007), while he de elopmen o he
elecommunica ion in as uc u e has a signi ican posi i e impac on he e iciency o go e n-
men (Do an e al., 2023). Thus, go e nmen s can also help p epa e he wo k o ce o he in eg a-
ion o AI in he public sec o . By in es ing in aining and eskilling p og ams, go e nmen s can
ensu e ha public sec o employees a e p epa ed o wo k wi h AI echnologies and manage he
changes associa ed wi h he au oma ion o ce ain asks (Duan e al., 2019). This can lead o a be -
e adap a ion o he wo k o ce o echnological de elopmen s and o a smoo he ansi ion in o
he digi al age. Wi h he de elopmen o ad anced language models using A i icial In elligence
(AI), he issue o AI eadiness becomes c i ical, and he quali y o he human ac o in ol ed in AI
depends la gely on he use o AI o pu poses uly use ul o humani y and wi hou associa ed
isks, especially hose ela ed o unp edic abili y and e hical conce ns.
Ano he impo an aspec o go e nmen spending is in es men in he in as uc u e
needed o suppo AI implemen a ion (Wang & Cui, 2022). This may include he de elop-
men o high-speed communica ion ne wo ks, da a cen e s and o he echnological esou ces
ha enable he e icien use o AI in public se ices and decision-making p ocesses. On he
o he hand, in es men s in in as uc u e, educa ion and esea ch and de elopmen can help
inc ease a coun y’s AI compe i i eness. S udies show ha go e nmen s in es ing in AI can
bene i om economic g ow h and imp o ed quali y o li e o ci izens (A n z e al., 2019).
Go e nmen in es men in AI can also os e in e na ional coope a ion (Milla d, 2017) and
pa ne ships be ween di e en go e nmen s and o ganiza ions, p omo ing a global and ha -
monized app oach o AI egula ions and s anda ds. This can lead o g ea e in e ope abili y
be ween AI sys ems used in di e en coun ies and o s eng hen collabo a ion o add ess
common challenges such as cybe secu i y and da a p o ec ion (Pan & Zhang, 2021).
Go e nmen spending on R&D in AI is ano he key issue o sus ainable economic de-
elopmen . The Eu opean Commission has made signi ican in es men s in he p ojec s o
A i icial In elligence (AI) esea ch and de elopmen i ms. This is done wi h he in en ion o
enhancing he p epa edness o go e nmen s o adop AI echnologies.
Gi en he p emise ha go e nmen al expendi u es a e di ec ed owa ds enhancing he
capabili y o go e nmen s o deploy a i icial in elligence, i is hypo hesized ha :
H2: The e is a posi i e e ec o go e nmen spending on AI go e nmen eadiness in he
Eu opean Union coun ies.
The Economic F eedom Index by The He i age Founda ion assesses economic libe y wo l-
dwide, o en used o examine he economic condi ions a o able o A i icial In elligence
(AI) g ow h. Na ions sco ing highe on his index ypically p esen a conduci e en i onmen
o AI adop ion, cha ac e ized by lowe axes, ewe egula ions, and be e p ope y igh s
p o ec ion.
Economic eedom is ins umen al o go e nmen s in AI eadiness, mainly by a ac ing
in es men and os e ing inno a ion. Coun ies wi h high sco es in he Index o Economic
F eedom a ac in es o s, hanks o s able economic condi ions and g ow h po en ial. This
en i onmen helps go e nmen s ga he necessa y capi al and expe ise o AI de elopmen
and implemen a ion (Wo ld Economic Fo um, 2023).
Mo eo e , coun ies wi h g ea e economic eedom usually ha e dynamic, compe i i e
ma ke s, spa king inno a ion and ad anced AI echnologies. In such ma ke s, companies
in es mo e in esea ch and de elopmen , s i ing o c ea e and ma ke op- ie AI p oduc s
(Ci ci & Du usu-Ci ci, 2022).
274 I. C. Iuga, A. Socol. Go e nmen A i icial In elligence eadiness and b ain d ain: in luencing ac o s and spa ial e ec s...
Ano he bene i o economic eedom o go e nmen AI eadiness is he p omo ion
o en ep eneu ial spi i and isk- aking. High-sco ing coun ies on he Index o Economic
F eedom o e a o able condi ions o en ep eneu ship, like easie c edi access and s ong
p ope y igh s. These ac o s encou age indi iduals and businesses o pu sue inno a i e
ideas, including AI- ela ed en u es. As a esul , go e nmen s in hese coun ies ha e access
o a con inuous s eam o inno a i e AI echnologies and applica ions (Le & Kim, 2020).
Economic eedom also acili a es he in eg a ion o AI echnologies wi hin go e nmen s.
High- anking coun ies on he Index o Economic F eedom ypically ha e e icien , s eamlin-
ed bu eauc a ic sys ems ha can easily adap o AI-d i en changes. These na ions ace ewe
bu eauc a ic hu dles o inno a ion, allowing go e nmen en i ies o expe imen wi h new AI
applica ions and echnologies mo e eely (Okulich-Kaza in e al., 2020).
In summa y, economic eedom signi ican ly in luences go e nmen AI eadiness. I d aws
in es men , s imula es inno a ion, os e s en ep eneu ship, and eases AI echnology in e-
g a ion wi hin go e nmen . Coun ies p io i izing economic libe y a e be e posi ioned o
le e age AI’s bene i s, hus eaping he ewa ds o his ans o ma i e echnology.
Conside ing hese, we can es a e:
H3: The e is a posi i e e ec o economic eedom on AI go e nmen eadiness in he Eu-
opean Union coun ies.
EU- unded p ojec s o R&D in AI a e ano he key elemen in suppo ing echnological
de elopmen and inno a ion. The Eu opean Union has in es ed massi ely in p og ams such
as Ho izon 2020 and he new Ho izon Eu ope p og am (2021–2027) o boos AI esea ch
and de elopmen (Eu opean Commission, 2021a). These ini ia i es o e inancial suppo and
collabo a i e p ospec s among esea che s, uni e si ies, and companies ac oss a ious Eu o-
pean coun ies, he eby ad ancing he sha ing o knowledge and he c ea ion o inno a i e
solu ions (Spence, 2021). A pe inen example is he AI4EU p ojec , unded by he EU unde
Ho izon 2020. I s i es o es ablish a Eu opean pla o m o he ad ancemen and u iliza ion
o AI, con ibu ing posi i ely o bo h he economy and socie y (AIoD Pla o m, 2019).
The subs an ial in es men by he Eu opean Commission in p ojec s ela ed o AI esea ch
and de elopmen i ms has no ably enhanced go e nmen eadiness o AI. These ini ia i es
ha e con ibu ed o be e go e nmen e iciency, inc eased ci izen engagemen and sa is ac-
ion, and augmen ed accoun abili y and anspa ency. As AI echnology con inues o e ol e,
hese ad an ages a e expec ed o become e en mo e dis inc i e, posi ioning AI as an inc eas-
ingly i al ins umen o go e nmen s globally. These p ojec s ha e p oduced conside able
posi i e e ec s, u nishing go e nmen s wi h he essen ial ools o le e age he po en ial o
AI and e olu ionize hei ope a ions.
A key ad an age o he Eu opean Commission’s unding in AI esea ch and de elopmen
i ms’ p ojec s lies in he imp o emen o go e nmen ope a ions’ e iciency and e icacy. U i-
lizing AI-powe ed ins umen s, go e nmen s can au oma e mundane asks and p ocedu es,
consequen ly cu ing ope a ional expenses, enhancing accu acy, and boos ing speed. Fo
example, AI can assis go e nmen s in swi ly and p ecisely p ocessing eno mous quan i ies
o da a, acili a ing supe io decision-making and mo e e icien esou ce alloca ion (Eu opean
Pa liamen , 2021a).
Ano he signi ican bene i o hese p ojec s is imp o ed ci izen engagemen and sa is-
ac ion. AI-d i en cha bo s and i ual assis an s can acili a e ci izens in accessing go e n-
men se ices wi h g ea e speed and e iciency, leading o educed wai ing imes and an
enhanced o e all expe ience. Addi ionally, AI-enabled ools can assis go e nmen s in mo e
Jou nal o Business Economics and Managemen , 2024, 25(2), 268–296 275
comp ehensi ely comp ehending he needs and p e e ences o hei ci izens (Ojo, 2019),
he eby allowing o he deli e y o mo e ailo ed se ices (Eu opean Pa liamen , 2021b).
The in es men s made by he Eu opean Commission in AI esea ch and de elopmen
i ms’ p ojec s ha e addi ionally con ibu ed o enhancing go e nmen accoun abili y and
anspa ency. AI-enabled ools can aid go e nmen s in o e seeing and assessing hei pe -
o mance, simpli ying he p ocess o pinpoin ing a eas ha need imp o emen and acking
p og ess o e ime. Mo eo e , AI can assis go e nmen s in iden i ying and a e ing aud
and co up ion, he eby gua an eeing ha public unds a e u ilized p ope ly (Eu opean Com-
mission, 2021b).
Conside ing he abo emen ioned, we can o mula e he las hypo hesis:
H4: The e is a posi i e e ec o EU- unded p ojec s o R&D on AI go e nmen eadiness in
he Eu opean Union coun ies.
3. Da a and me hodology
Ou s udy analyzes whe he go e nmen s’ eadiness o a i icial in elligence AI in he Eu o-
pean Union (EU) coun ies depends on “b ain d ain”, using Go e nmen A i icial In elligence
Readiness Index as dependen a iable and he Human ligh and b ain d ain as he majo
independen p oxy. Se e al economic, esea ch and de elopmen and go e nance ac o s a e
conside ed con ol a iables, based on he esul s iden i ied in he p e ious s udied li e a u e.
To con ol he mac oeconomic condi ions, he s udy use Go e nmen Expendi u e, G oss
Domes ic P oduc and Economic F eedom, while o con ol go e nance and educa ion in AI
ield, wo p oxies a e employed: Go e nmen Pilla o AI Index and AI in Uni e si y Bachelo ’s
P og ams (p opo ion o p og ams wi h AI con en in he o al numbe o p og ams). Also,
he s udy conside s he a iable ep esen ed by AI R&D Fi ms’ P ojec s ounded by Eu ope-
an Commission (pe cen age o he o al numbe o AI R&D playe s inanced) as a p oxy o
cap u e Resea ch and De elopmen ac i i ies in AI.
The panel comp ises he Eu opean Union coun ies, excep o Cyp us, Mal a and I e-
land, which a e excluded om he s udied sample, gi en hei geog aphical s a us wi hou
EU neighbo ing coun ies (bases on spa ial analysis me hods ha do no allow he analysis
o hose s a es ha do no ha e common bo de s wi h he es o he s a es in he chosen
sample).
Table 1 p esen s he desc ip ion o he a iables and da a sou ces om which he in o -
ma ion was ga he ed.
To bes o ou knowledge, a egional o global le el, he sca ci y o indica o s o cap u e
he p epa a ion o go e nmen s in he implemen a ion o a i icial in elligence is ob ious,
because he e a e limi ed ini ia i es o de elop such indica o s. We a e in e es ed in cap u -
ing he si ua ion o AI implemen a ion in go e nmen in he pe iod as close o he p esen
as possible, so we use a yea window which e e s o he yea 2022, and o he es o he
a iables we ely on he mos ecen da a collec ion a ailable, published in 2022 and ela ed
o he p e ious yea o yea s, depending on he collec ion c i e ia o hose a iables.
The issue o p epa ing go e nmen s o AI deploymen is a ely s udied in he li e a u e,
much less i s connec ion o b ain d ain, gi en he no el y o he subjec and he incipien
conce ns o scien i ic communi ies o analyze uncon en ional pe spec i es o AI implemen a-
ion by go e nmen s. As a as he au ho ’s knowledge is conce ned, no s udy has explo ed
such a opic om he pe spec i e o he Eu opean Union and he a iables chosen by us.
282 I. C. Iuga, A. Socol. Go e nmen A i icial In elligence eadiness and b ain d ain: in luencing ac o s and spa ial e ec s...
o Ande son-Rubin and no signi ican o he Basman es ). Also, in he 2SLS model,
we pe o m es s o s udy he explana o y powe o he ins umen s, whose ou comes
con i m ha addi ional ins umen s (lnGDP, GOVERNANCE and AI BACHELOR) ha e signi -
ican explana o y powe o AI IN GOVERNANCE a e con ol o he e ec o he es o
he independen a iables o he model (F s a is ic signi ican a 1% le el). The minimum
eigen alue s a is ic as a es o weak ins umen s compa ed wi h c i ical alues o 2SLS
ela i e bias deno es ha ou ins umen s a e no weak ( he F s a is ic exceeds he c i ical
alue i we we e willing ole a e a 10% ela i e bias).
As a human p oxy is conce ned, BRAIN DRAIN has a signi ican nega i e in luence on AI
IN GOVERNMENT, which shows ha he inc ease in he lows o human ligh and b ain d ain
om a coun y ha is losing human capi al we a e wi nessing a educ ion in he go e nmen ’s
abili y o implemen and manage a i icial in elligence in public se ices o ci izens. The e
a e only ci cums an ial indica ions wi hin he ci ed li e a u e sugges ing a nega i e e ec o
b ain d ain on go e nmen AI eadiness. Ou indings a e consis en wi h his body o li e -
a u e (Czaika & de Haas, 2015; Sia , 2013; B ynjol sson & McA ee, 2014; Dignum, 2019) and
alida e he main esea ch hypo hesis H1 ha b ain d ain nega i ely a ec s AI in go e nmen .
The b ain d ain phenomenon can ha e a nega i e impac on a coun y’s go e nmen ’s AI
eadiness, o se e al easons:
1. Sho age o AI-skilled human esou ces can slow down o e en s op he de elopmen
and implemen a ion o AI-based solu ions in he public sec o (Vicsek, 2021);
2. B ain d ain can nega i ely a ec he le el o inno a ion in a coun y (Spence, 2021)
(when AI expe s lea e he coun y, hey ake hei knowledge and expe ience wi h
hem, which can lead o a dec ease in scien i ic and echnological ad ances in AI);
3. B ain d ain can lead o a loss o in es men in educa ion and aining: go e nmen s
in es in he educa ion and aining o AI specialis s, wi h he hope ha hey will con-
ibu e o he coun y’s economic de elopmen and g ow h. When hese specialis s
choose o wo k ab oad, he go e nmen loses he bene i s o i s in es men s, while
he hos coun ies bene i om hese alen s;
4. B ain d ain can c ea e majo di icul ies in in e na ional collabo a ion: expe s who
lea e hei home coun y may be less in e es ed in wo king wi h he go e nmen and
ins i u ions in hei coun y o o igin, which can make i di icul o access in e na ional
knowledge and esou ce ne wo ks;
5. B ain d ain can weaken abili y o compe e globally: as o he coun ies a ac alen
in o AI, hey can become mo e echnologically and economically compe i i e. A he
same ime, he coun y o o igin could be le behind in he global AI ace.
Adop ing s a egies can help go e nmen s coun e he nega i e e ec s o b ain d ain on
go e nmen AI aining and de elop an en i onmen conduci e o g owing AI expe ise and
skills in he coun y. F om he mul i ude o s a egies ha can be applied by go e nmen s,
we p esen he mos impo an ones:
1. Enhancing educa ion sys ems: In es ing in academic and esea ch ins i u ions o deli -
e supe io AI p og ams, de eloping AI aining and esea ch ini ia i es, and os e ing
pa ne ships be ween uni e si ies, businesses, and go e nmen al en i ies. P omo ing
STEM educa ion: Encou aging in e es in AI om an ea ly school le el h ough em-
phasis on science, echnology, enginee ing, and ma hema ics.
2. P o iding inancial incen i es and suppo : Alloca ing go e nmen unds o AI e-
sea ch and de elopmen and es ablishing a ax- iendly en i onmen o companies
de eloping and applying AI echnologies.
Jou nal o Business Economics and Managemen , 2024, 25(2), 268–296 283
3. C ea ing a conduci e en i onmen o inno a ion: Encou aging collabo a ion be ween
public and p i a e sec o s o ackle socie al and go e nmen al issues ia AI, and a-
cili a ing access o essen ial in as uc u e and esou ces o AI solu ion de elopmen
and implemen a ion (B ed , 2019).
4. A ac ing and e aining o AI alen : implemen ing a o able mig a ion policies o
a ac and e ain o eign AI alen ; p o iding compe i i e and ad an ageous ca ee
oppo uni ies in he public sec o o AI specialis s. E en mo e, he policy o inc easing
wages and bene i s can also be applied (go e nmen s can inc ease wages and bene i s
o public sec o AI p o essionals o encou age e aining and a ac ing alen . This
s a egy could also include he de elopmen o bonus schemes and o he bene i s
ha would a ac and e ain AI p o essionals in he coun y) (Ag awal e al., 2019).
5. In e na ional coope a ion and knowledge exchange: pa icipa ion in in e na ional AI
alliances, pa ne ships and p ojec s o bene i om he expe ise and esou ces o
o he coun ies; p omo ing academic and p o essional exchanges in he ield o AI
o acili a e he ans e o knowledge and good p ac ice; suppo ing coope a ion
be ween esea ch and inno a ion ins i u ions in di e en coun ies o wo k oge he
in he de elopmen o AI solu ions.
6. AI in as uc u e de elopmen : go e nmen s can de elop he in as uc u e needed o
suppo AI de elopmen , such as da a cen e s, communica ion ne wo ks, and cloud
compu ing in as uc u e (Duan e al., 2019). This s a egy could also include in es -
men s in he coun y’s digi al in as uc u e, including connec i i y and he de elop-
men o ad anced digi al echnologies.
Rega ding he con ol a iables o he models, he con ibu ion o GOVERNMENT EX-
PENDITURE in s eng hening he capaci y o go e nmen s o implemen a i icial in elligence
is con i med and demons a es a posi i e ela ionship wi h dependen a iable: he highe
he olume o go e nmen spending, he highe he go e nmen ’s eadiness in a i icial in-
elligence. The esul s con i m he gene al assump ions om p e ious li e a u e (Wang &
Cui, 2022; B ed , 2019; Duan e al., 2019; Bose e al., 2007) and he eso s o such an in-
e dependence a e ela ed o he inno a i e and ela i ely expensi e echnical cha ac e o
he s a e’s in es men s in he a i icial in elligence in as uc u e and so wa e o he public
se ices. Also, he pe sonnel needed o de elop and manage in o ma ic applica ions and
in as uc u e based on a i icial in elligence leads o signi ican go e nmen spending. Gi en
he p emise ha go e nmen al expendi u es a e di ec ed owa ds enhancing he capabili y o
go e nmen s o deploy a i icial in elligence, he ou comes o he s udy empi ically alida e
and subs an ia e Resea ch Hypo hesis H2, which posi s a posi i e e ec o such spending on
AI go e nmen eadiness in he Eu opean Union coun ies.
ECONOMIC FREEDOM is posi i ely signi ican ly associa ed wi h he go e nmen ’s abili y o
implemen a i icial in elligence (Ci ci & Du usu-Ci ci, 2022; Le & Kim, 2020; Okulich-Kaza in
e al., 2020). The mo e economic eedom is expe ienced in EU coun ies, he mo e i c ea es
p emises o go e nmen s o adop a i icial in elligence. D awing om he empi ical indings
o he s udy, Resea ch Hypo hesis H3, which posi s ha he e is a posi i e e ec o economic
eedom on AI go e nmen eadiness in he Eu opean Union coun ies, has been alida ed.
The inancing by he Eu opean Commission o esea ch & de elopmen i ms o a i i-
cial in elligence p ojec s is also ound o be a signi ican de e minan o inc easing a i icial
in elligence in public adminis a ion (Spence, 2021). The highe he pe cen age o esea ch &
de elopmen companies ha bene i om inancing a i icial in elligence p ojec s, he mo e
na ional go e nmen s bene i om scien i ic esea ch esul s in a i icial in elligence, which
284 I. C. Iuga, A. Socol. Go e nmen A i icial In elligence eadiness and b ain d ain: in luencing ac o s and spa ial e ec s...
can be de eloped and implemen ed a go e nmen le el. Upon igo ous examina ion o he
da a, Resea ch Hypo hesis H4, which pos ula es a posi i e e ec o EU- unded p ojec s o
R&D on AI go e nmen eadiness in he Eu opean Union coun ies, has been alida ed.
In he ins umen al- a iables eg essions, GOVERNMENT EXPENDITURE is conside ed he
ins umen ed a iable, whose con igu a ed ins umen s a e lnGDP, GOVERNANCE and AI
BACHELOR. Employing GDP as an indica o o assess he in luence o AI on he economy has
been bene icial o gauging go e nmen al AI eadiness. GDP, a p e alen me ic o economic
ac i i y, when used o measu e AI’s impac , unde sco es he echnology’s signi icance o
economic expansion. By moni o ing AI’s con ibu ion o GDP, go e nmen s can achie e a
mo e p o ound comp ehension o AI’s economic bene i s and iden i y he sec o s whe e AI
can make he mos subs an ial impac (Piasecki e al., 2021).
The de elopmen o Go e nmen Pilla o AI Index has been a posi i e de elopmen o
go e nmen AI eadiness. The index is a ool o assessing he eadiness o go e nmen s o
AI adop ion, and i includes a ange o me ics ela ed o AI policies, egula ions, and in a-
s uc u e. I p o ides he assessmen o ision, go e nance and e hics, digi al capaci y, and
adap abili y o go e nmen s in AI implemen a ion.
By using he index o benchma k hei AI eadiness, go e nmen s can iden i y a eas o
imp o emen and de elop s a egies o suppo he g ow h o AI in hei coun ies.
The Go e nmen Pilla o AI Index also p o ides a pla o m o in e na ional compa ison,
allowing go e nmen s o assess hei AI eadiness in ela ion o o he coun ies. This can help
o p omo e compe i ion and encou age go e nmen s o adop policies and s a egies ha
suppo he g ow h o AI in hei coun ies.
As AI’s ole in shaping he u u e o wo k expands, i s in eg a ion in o Uni e si y Bache-
lo ’s P og ammes has gained impo ance. Go e nmen s globally acknowledge he signi icance
o AI eadiness and de eloping AI s a egies o sus ainable economic g ow h. The use o GDP
and AI in Uni e si y Bachelo ’s P og ammes, as well as he de elopmen o Go e nmen Pilla
o AI Index, ha e bo h had a posi i e impac on go e nmen AI eadiness.
The d i e o in eg a e AI in o Uni e si y Bachelo ’s P og ammes s ems om he inc eas-
ing demand o AI skills in he job ma ke . The applica ion o AI ools is quickly becoming
an in eg al pa o nume ous indus ies, including inance, heal hca e, and manu ac u ing.
Uni e si ies, by inco po a ing AI in hei p og ammes, equip s uden s wi h i al skills o he
u u e wo kplace (Vicsek, 2021).
Mo eo e , his in eg a ion uels inno a ion and economic g ow h. AI is a po en inno-
a ion ool, and he eme gence o new AI echnologies could s imula e new indus ies and
jobs. Uni e si ies, by eaching s uden s he skills needed o c ea e and apply AI solu ions, spu
economic g ow h and pa e he way o new oppo uni ies o g adua es.
In summa y, he use o GDP as an AI impac measu e, he c ea ion o he Go e nmen
Pilla o he AI Index, and he in eg a ion o AI in o Uni e si y Bachelo ’s P og ammes ha e
posi i ely in luenced go e nmen AI eadiness. By acing AI’s economic bene i s, p o iding
a gauge o go e nmen AI eadiness, and p epa ing s uden s wi h essen ial skills o he
u u e wo kplace, hese ini ia i es bols e global go e nmen s’ capaci y o le e age he op-
po uni ies AI o e s.
4.2. Clus e analysis
To es ablish g oups o EU coun ies and disco e hei mu ual simila i ies based on he se
o independen a iables associa ed wi h he go e nmen a i icial in elligence eadiness, we
Jou nal o Business Economics and Managemen , 2024, 25(2), 268–296 285
ca ied ou he clus e analysis o he yea 2022. EU coun ies a e cha ac e ized by dispa i ies
in socio-economic de elopmen and unequal na ional e olu ion, especially as he s a es ha
joined he Eu opean Union in 2004 and 2007 (Bulga ia, Czech Republic, Es onia, Hunga y,
La ia, Li huania, Poland, Romania, Slo akia and Slo enia) expe ienced cen alized communis
economies un il he ea ly 1990s, ha e since passed h ough di icul pe iods o ansi ion
o he ma ke economy and ha e had o o e come s uc u al di e ences and gaps in hei
economies compa ed o hose o he de eloped coun ies o he Eu opean Union. In addi ion,
adi ionally, na ional inequali ies also de i e om he di e en size o s a es, disc epancies
in equipmen and public echnological in as uc u e. The de e mina ion o homogeneous
clus e s o s a es in e ms o AI implemen a ion in go e nmen is qui e challenging gi en he
ob ious condi ions o he e ogenei y exposed.
The clus e s associa ed wi h he go e nmen a i icial in elligence eadiness index we e
de eloped h ough wo ca ego ies o hie a chical clus e ing algo i hm (comple e linkage ha
p oduces spa ially compac clus e s and Wa d’s linkage whe e g oups we e joined o maxi-
mize an e o -sum-o -squa es objec i e unc ion), which we e applied o es ablish he g oups
o homogeneous s a es and which c ea es a hie a chy o clus e s, based on dissimila i y meas-
u e in compa ing wo obse a ions. The sui able numbe o clus e s was es ablished based
on clus e s opping ules (Duda-Ha pseudo-T-squa ed small alues and Calinski-Ha abasz
pseudo-F la ge alues, which bo h indica ed ou clus e s), as well as om he s udy o he
dend og am and g aphic ep esen a ion.
By applying he comple e linkage and Wa d’s me hods o yea 2022, ou coun ies clus-
e s based on hei simila i ies a e ob ained (Table 5). The syn he ic desc ip ion o he cha ac-
e is ics o he coun ies in each clus e , acco ding o he s udied a iables, con i ms ha he
deg ee o implemen a ion o a i icial in elligence in public se ices is in e sely p opo ionally
wi h he b ain d ain phenomenon. Also, he posi i e associa ions o he dependen a iable
wi h spending expenses, economic eedom and esea ch and de elopmen p ojec s unded
by he Eu opean Commission o AI companies a e con i med.
Table 5. Clus e s by AI IN GOVERNMENT – comple e linkage and Wa d’s me hods
G oup o coun y AI IN GOVERNMENT dimension
Numbe o
clus e –
comple e
linkage
me hod
Numbe
o clus-
e – Wa d’s
me hod
Coun ies wi h he a e age highes deg ee o go e nmen a i icial in elligence
Czechia, Finland,
Ne he lands,
Luxembou g,
Aus ia, Denma k,
Es onia, Sweeden
High o medium (in e ms o AI IN GOVERNMENT)
Medium o low (in e ms o BRAIN DRAIN)
High (in e ms o ECONOMIC FREEDOM and AI R&D
FIRMS EC FUNDED PROJECTS)
High o medium and low (in e ms o GOVERNMENT
EXPENDITURES)
Clus e 1 Clus e 1
Coun ies wi h high le el o go e nmen a i icial in elligence
F ance, Ge many High o medium (in e ms o AI IN GOVERNMENT)
Medium (in e ms o BRAIN DRAIN, ECONOMIC
FREEDOM and AI R&D FIRMS EC FUNDED PROJECTS)
High (in e ms o GOVERNMENT EXPENDITURES)
Clus e 2 Clus e 3
286 I. C. Iuga, A. Socol. Go e nmen A i icial In elligence eadiness and b ain d ain: in luencing ac o s and spa ial e ec s...
G oup o coun y AI IN GOVERNMENT dimension
Numbe o
clus e –
comple e
linkage
me hod
Numbe
o clus-
e – Wa d’s
me hod
Coun ies wi h mode a e le el o go e nmen a i icial in elligence
Belgium,
Hunga y,
I aly, G eece,
Po ugal,
Slo enia, Spain
Medium (in e ms o AI IN GOVERNMENT)
Medium o low (in e ms o BRAIN DRAIN)
High o medium (in e ms o AI R&D FIRMS EC
FUNDED PROJECTS)
High o medium and low (in e ms o GOVERNMENT
EXPENDITURES)
Medium o low (in e ms o ECONOMIC FREEDOM)
Clus e 3 Clus e 2
Coun ies wi h low deg ee o go e nmen a i icial in elligence
Romania,
Slo akia, C oa ia,
La ia,
Poland, Bulga ia,
Li huania
High o medium (in e ms o AI IN GOVERNMENT)
High (in e ms o BRAIN DRAIN)
Medium o low (in e ms o GOVERNMENT
EXPENDITURES, ECONOMIC FREEDOM)
Low (in e ms o
AI R&D FIRMS EC FUNDED PROJECTS)
Clus e 4 Clus e 4
The geog aphical ep esen a ion o he coun ies based on he main p oxy BRAIN DRAIN
compa ed wi h he clus e ’s ep esen a ion in comple e linkage me hods a e illus a ed in
Figu e 1.
The lowes alues o co e explana o y a iable (BRAIN DRAIN) a e in Romania, Bulga ia,
C oa ia, Slo akia, Poland, La ia, Li huania and Es onia. This g oup o s a es o e laps (wi h
he excep ion o Es onia) wi h he las clus e ob ained, clus e 4, de e mined on he basis o
clus e analysis and con aining s a es wi h low deg ee o go e nmen a i icial in elligence.
The simila ype o co espondence be ween he wo maps (Figu e 1) is ound o he i s
h ee clus e s, in he sense ha he coun ies in ca ego ies 2, 3 and 4 a e he dec easing
olume o he b ain d ain a e in dis inc clus e s and o he en i e se o da a analysis by
he clus e me hod, in which all a iables we e conside ed in he AI go e nmen es ima ion
(Figu e 1b and Table 6). Fo example, he coun ies wi h he lowes b ain d ain alues (Finland,
Figu e 1. Geog aphical ep esen a ion o he EU (24) coun ies based on: a – b ain d ain;
b – Clus e s in comple e linkage me hod
a) b)
End o Table 5
Jou nal o Business Economics and Managemen , 2024, 25(2), 268–296 287
Ne he lands, Luxembou g, Denma k, and Sweden) a e included in clus e 1 wi h he a e -
age highes deg ee o go e nmen a i icial in elligence. S a es such as Czechia and Aus ia,
placed in clus e 1 and ha ing medium b ain d ain alues a e in he i s pa o he s a es
s udied om he pe spec i e o adop ing AI in go e nmen . Es onia’s case is a special one, in
he sense ha al hough i p esen s high alues o b ain d ain mig a ion lows, i has de el-
oped and consolida ed AI in go e nmen , h ough a ge ed and success ul na ional measu es.
Ge many and F ance (clus e 2 in comple e linkage me hod) p esen medium alues in e ms
o b ain d ain and high le el o go e nmen a i icial in elligence, while he es o coun ies
(clus e 3 – Belgium, Hunga y, I aly, G eece, Po ugal, Slo enia, Spain) shows a e age o low
b ain d ain alues, as well as a e age alues o AI in go e nmen . The analyzed geog aphical
dis ibu ion e eals si ua ions in which immedia ely neighbo ing s a es ha e he same ype o
beha io om he pe spec i e o he clus e s o belonging and he a ia ion o he a iables
s udied by he clus e analysis, which adds p elimina y signals o in luences be ween s a es,
as we will s udy in he ollowing h ough spa ial analysis.
To de e mine i he selec ed independen a iables ha e he po en ial o explain he e o-
lu ion o he deg ee o implemen a ion o he go e nmen a i icial in elligence, we es ima e
a clus e mul i a ia e eg ession model (Table 6). The esul s display high a es o explana ion
o he chosen model o each independen a iable (R-squa ed), as well as s a is ically sig-
ni ican ou comes o he F s a is ic, which indica es ha he model has p edic i e capabili y.
Table 6. Clus e s analysis o AI IN GOVERNMENT (sou ce: au ho 's calcula ions using STATA)
AI IN GOVERNMENT F R-squa ed
BRAIN DRAIN 13.596*** 0.671
GOVERNMENT EXPENDITURE 29.465*** 0.815
ECONOMIC FREEDOM 10.468*** 0.610
AI R&D FIRMS EC FUNDED PROJECTS 21.356*** 0.762
No e: * Signi ican a 10%; ** Signi ican a 5%; *** Signi ican a 1%.
4.3. Spa ial analysis
A e iden i ying he ela ionship be ween he a iables s udied by he ins umen al- a iables
eg ession me hod and clus e coun ies analysis, ano he aim o he s udy is o ind he
spa ial pa e n o he da a, mo e p ecisely whe he AI IN GOVERNMENT and i s de e minan s
display agglome a ion o high alues wi h high alues be ween coun ies, low alues wi h low
alues o high alues is g ouped nex o low alues.
To pe o m he spa ial analysis, we ini ially design a spa ial weigh ma ix in he GeoDa
so wa e (Anselin e al., 2006), which de ines he neighbo hoods be ween s a es, in a bina y
o ma , in which i he e is a common bo de be ween coun ies he alue is 1 and o h-
e wise 0. As he weigh ma ix canno e lec he speci ic ela ionships be ween coun ies
(human, economic o go e nance in ou case), we de elop he nes ed spa ial weigh ma ix
ob ained by mul iplying he adjacen ma ix wi h he weigh ma ix o human, economic o
go e nance dis ance om ou da a. We use he nes ed s anda dized ma ix, which can e lec
he spa ial cha ac e is ics o neighbo ing coun ies, as well as he deg ee o in e dependence
be ween coun ies (Elho s , 2017; Wang e al., 2023).
Based on nes ed ma ix, we es ima e he global Mo an index I and ob ain posi i e alues o
i and qui e close o 1 o he explained and co e explana o y a iables (AI IN GOVERNMENT
288 I. C. Iuga, A. Socol. Go e nmen A i icial In elligence eadiness and b ain d ain: in luencing ac o s and spa ial e ec s...
and BRAIN DRAIN) which indica es ha hese a iables ha e a subs an ial posi i e spa ial au-
oco ela ion, and he da a is g ouped on he p inciple high-high o low-low (Table 7). Values
a e also posi i e, bu close o 0 a e ob ained o con ol a iables and show a lowe deg ee
o spa ial g ouping o obse a ions. O e all, da a p esen s signi ican spa ial co ela ion and
is sui able o hei use in spa ial models.
Table 7. Global Mo an’s I spa ial co ela ion o he selec ed a iable, yea 2022
Va iable(s) Mo an’s I p- alue z- alue
AI IN GOVERNMENT 0.479 0.002 2.818
BRAIN DRAIN 0.699 0.000 3.930
GOVERNMENT EXPENDITURE 0.193 0.071 1.469
ECONOMIC FREEDOM 0.258 0.054 1.612
AI R&D FIRMS EC FUNDED PROJECTS 0.267 0.048 1.666
No e: * Signi ican a 10%; ** Signi ican a 5%; *** Signi ican a 1%. S anda d e o s a e in ound b acke s.
Following he con i ma ion o he exis ence o spa ial co ela ions, he spa ial e o e-
g ession model is examined. The obus ness check o he model shows ha he lambda
pa ame e is signi ican (as a e he Wald and Lag ange mul iplie es s), and he alues o
a iance a io and squa ed co ela ion ensu e a conside able p edic o dependen a iable
h ough explana o y indica o s.
F om he empi ical esul s (Table 8) we can ind ha bo h co e explana o y a iable and
con ol a iables p esen s a is ically signi ican alues o spa ial co ela ion coe icien s and
explain go e nmen AI om spa ial pe spec i e be ween coun ies. The inc easing BRAIN
Table 8. Spa ial E o Model eg ession esul s (sou ce: au ho s’ calcula ions using STATA)
AI IN GOVERNMENT SEM
BRAIN DRAIN –1.826***
(0.633)
GOVERNMENT EXPENDITURE 0.008**
(0.003)
ECONOMIC FREEDOM 0.703***
(0.108)
AI R&D FIRMS EC FUNDED PROJECTS 22.039*
(11.850)
Cons an 5.595
(14.669)
Va iance a io 0.843
Squa ed co ela ion 0.825
Lambda –0.466***
(0.127)
Wald es 13.378***
Lag ange mul iplie LM es 4.594**
No e: * Signi ican a 10%; ** Signi ican a 5%; *** Signi ican a 1%. Robus s anda d e o s a e in ound b acke s.
Jou nal o Business Economics and Managemen , 2024, 25(2), 268–296 289
DRAIN in neighbo ing coun ies has a spa ial spillo e e ec s o nega i e ex e nali y on he
home coun y’s AI IN GOVERNMENT. Also, he imp o emen o mac oeconomic condi ions
(GOVERNMENT EXPENDITURE and ECONOMIC FREEDOM), as well as hose o inancing AI
R&D (AI R&D FIRMS EC FUNDED PROJECTS) in neighbo ing coun ies a e esponsible o
inc easing he capaci y o go e nmen s o manage AI adop ion.
Spa ial analysis o he in luence o b ain d ain on AI in go e nmen highligh s egional
and global inequali ies in he de elopmen and deploymen o AI echnologies in he public
sec o . Add essing hese inequali ies equi es conce ed e o s by go e nmen s, academia
and he p i a e sec o o de elop a s ong alen base in AI, encou age in e na ional col-
labo a ion and suppo he de elopmen o local AI solu ions ailo ed o he speci ic needs
o di e en go e nmen s.
Spa ial analysis o he in luence o he b ain d ain on AI in go e nmen highligh s egional
and global dispa i ies in he de elopmen and implemen a ion o a i icial in elligence in he
public sec o . B ain d ain e e s o he mig a ion o highly quali ied specialis s, including AI
expe s, om coun ies o o igin o des ina ions wi h be e oppo uni ies, such as de eloped
coun ies (Kapu & McHale, 2005). This alen mig a ion can ha e a nega i e impac on go -
e nmen s’ abili y o de elop and deploy AI solu ions.
Regional imbalances in AI de elopmen can c i ically a ec go e nmen e iciency in pub-
lic se ice deli e y and policy o ma ion (Wi z e al., 2020). Go e nmen s wi h AI alen access
and he esou ces o ealize AI solu ions a e be e equipped o ackle complex social and
economic p oblems, mee ing ci izen needs (Mikhaylo e al., 2018). Con e sely, go e nmen s
ailing o a ac and e ain AI specialis s could see hei public se ice and policy manage-
men e iciency decline (Zajko, 2022).
The b ain d ain can exace ba e inequali ies be ween coun ies in access o inno a ions
and ad anced echnologies in AI. Coun ies ha manage o a ac and e ain alen in AI can
bene i om in es men in R&D and accele a ed economic g ow h (Singh & K ishna, 2018).
The concen a ion o alen in AI in ce ain egions o coun ies can ampli y di e ences in AI
de elopmen and adop ion in go e nmen s a ound he wo ld. De eloped coun ies, especially
Wes e n Eu opean coun ies, ha e managed o a ac many AI specialis s, s eng hening hei
AI posi ions. In con as , de eloping coun ies and local go e nmen s may ace di icul ies in
a ac ing and e aining AI alen (Bane jee & Du lo, 2019).
A he same ime, coun ies a ec ed by he b ain d ain may lag behind in he echnologi-
cal ace, which may limi hei abili y o add ess social and economic p oblems and espond
o ci izens’ needs (Sab y, 2021).
To add ess hese spa ial inequali ies and encou age a mo e equi able de elopmen o
AI in he public sec o , go e nmen s need o conside se e al s a egies. Fi s , in es men in
educa ion and esea ch in AI should be encou aged and suppo ed a local and egional le el
(Bas i & Box, 2010). This can help de elop a solid alen base in AI and c ea e employmen
oppo uni ies in he public sec o (Yamashi a e al., 2021). Second, in e na ional AI collabo a-
ion can be c ucial o comba he b ain d ain and ensu e ha o me communis EU coun-
ies ha e access o cu ing-edge knowledge and echnologies (Be ge , 2022). Pa ne ships
be ween coun ies and ins i u ions can acili a e he ans e o echnology and knowledge
and con ibu e o he de elopmen o AI solu ions ailo ed o he speci ic needs o di e en
go e nmen s (Qin e al., 2023). Thi d, go e nmen s can encou age he de elopmen o lo-
cal AI solu ions and suppo small and medium-sized en e p ises (SMEs) specializing in AI.
C ea ing an inno a i e AI ecosys em can a ac local and in e na ional alen and help build
go e nmen s’ capaci ies o implemen AI in he public sec o (Dje al e al., 2022).
290 I. C. Iuga, A. Socol. Go e nmen A i icial In elligence eadiness and b ain d ain: in luencing ac o s and spa ial e ec s...
The indings lead o a ious in e ences and ecommenda ions. Fo public adminis a ions,
AI ex ends beyond being me ely ins umen al; i ’s a ans o ma i e ca alys wi h po en ial o
eshape go e nance, policy o ma ion, and public se ice p o ision. AI can help go e nmen s
gene a e accu a e p edic ions and simula e complex sys ems o expe imen wi h di e en policy
op ions (Ma ge s & Do oban u, 2019). Hence, i ’s c ucial o public se an s o keep pace wi h
AI ad ancemen s and comp ehend i s applicabili y o socie al wel a e (Reis e al., 2019).
Mo eo e , AI’s ole in go e nmen unde sco es he necessi y o obus collabo a ions
amongs public ins i u ions, p i a e en e p ises, academic bodies, and ci il socie y o ganiza-
ions. Such pa ne ships encou age knowledge exchange, esou ce sha ing, and expe ise
pooling, which a e i al o c a ing AI sys ems ailo ed o go e nmen al needs and goals.
I is impo an o no e ha b ain d ain impedes go e nmen AI de elopmen , as he
emig a ion o AI p o essionals slows AI p og ess in public se ices, educing expe ise and
inno a ion. This esul s in economic losses o go e nmen s, especially in esou ce-limi ed
na ions, and hampe s in e na ional AI collabo a ion due o less coope a ion om eloca ed
expe s. Such na ions isk alling behind in he global AI ace, while hose e aining alen
could ad ance echnologically and economically. Go e nmen spending is key o AI eadiness,
and s able economic policies a o AI adop ion. The Eu opean Commission’s AI R&D in es -
men unde sco es he need o ongoing inancial suppo in his ield.
Finally, add essing b ain d ain is i al o a na ion’s AI u u e, economic g ow h, and in-
no a ion. Go e nmen s mus ackle hese challenges wi h p oac i e s a egies, os e ing an
en i onmen conduci e o AI p og ess o he bene i o all ci izens.
5. Conclusions
Gi en he in e sec ion be ween A i icial In elligence (AI) and b ain d ain, his s udy ga ne s
b oad public in e es by p o iding aluable insigh s, enhancing unde s anding o AI’s impac
on socie y, economy, and en i onmen and highligh ing he need o go e nmen s o egu-
la e and implemen AI ac oss a ious sec o s.
This s udy o e s nuanced unde s andings o he ac o s in luencing AI go e nmen eadi-
ness wi hin he Eu opean Union. Speci ically, i del es in o he po en ial nega i e epe cus-
sions o B ain D ain (RQ1) and he po en ial posi i e implica ions o EU- unded p ojec s o
R&D, economic eedom, and go e nmen spending (RQ2) on AI p epa edness. Mo eo e ,
he esea ch p o ides c ucial insigh s in o he in e play be ween b ain d ain and a coun y’s
in luence on i s neighbo ing s a es’ AI adop ion eadiness (RQ3). These indings elucida e he
signi icance o unde s anding bo h in e nal and ex e nal dynamics o enhance AI eadiness
and in eg a ion a he go e nmen al le el in he Eu opean egion.
Ou indings unequi ocally indica e a nega i e co ela ion be ween b ain d ain and AI
eadiness in EU go e nmen s in 2022. This sugges s ha as b ain d ain in ensi ies, go e n-
men s’ p epa edness o in eg a e AI diminishes. This co ela ion, coupled wi h he insigh s
ga ne ed om he spillo e e ec s and he po en ial policy in e en ions, o e s a comp e-
hensi e oadmap o EU na ions o s a egically add ess he challenges posed by b ain d ain
in he ealm o AI eadiness.
Se e al implica ions a ise om ou indings:
De imen al E ec s o B ain D ain: B ain d ain signi ican ly unde mines go e nmen AI
eadiness. Skilled AI p o essionals mig a ing ab oad impede he p og ess and applica ion o
ad anced AI in public se ices, con i ming Hypo hesis H1 and suppo ing exis ing esea ch
indings.
Jou nal o Business Economics and Managemen , 2024, 25(2), 268–296 291
Inno a ion Lag: The mig a ion o AI alen leads no only o a loss o expe ise bu also
diminishes he po en ial o echnological inno a ion wi hin hei home coun ies.
Economic Implica ions: When AI expe s lea e, hei home coun ies lose he in es men s
made in hei educa ion and aining. This exace ba es economic di icul ies, especially in
esou ce-cons ained na ions.
Challenges in In e na ional Collabo a ion: The b ain d ain phenomenon can ad e sely
a ec in e na ional AI collabo a ions. Expe s eloca ing ab oad migh become less willing o
engage wi h hei home coun ies, esul ing in los oppo uni ies in in e na ional AI p ojec s.
Global AI Race: Coun ies se e ely a ec ed by b ain d ain isk alling behind in he in e -
na ional compe i ion o AI ad ancemen . In con as , hose e aining o a ac ing AI alen
could see signi ican echnological and economic g ow h.
P ominence o Go e nmen Expendi u e: The e is a no able co ela ion be ween go e n-
men in es men and AI eadiness. Adequa e go e nmen unding can subs an ially imp o e
he public sec o ’s AI capabili ies.
The Role o Economic F eedom: An enabling economic en i onmen is c ucial o he ad-
op ion o AI in go e nmen . Coun ies wi h g ea e economic eedom ypically ha e an
easie pa h o AI in eg a ion, emphasizing he impo ance o s able and suppo i e economic
policies.
Eu opean Commission’s AI R&D In es men : The Eu opean Commission’s unding in AI
esea ch and de elopmen has been pi o al. This in es men is essen ial o ad ancing AI
eadiness, highligh ing he impo ance o ongoing inancial suppo in his ield.
Add essing b ain d ain is c ucial o a na ion’s u u e in he AI a ena, ensu ing economic
g ow h and inno a ion. Go e nmen s need s a egies o e ain alen and c ea e AI- iendly
en i onmen s, bene i ing ci izens. Howe e , ou s udy aces limi a ions, including da a a ail-
abili y cons ain s. Despi e using eliable sou ces, da a gaps and disc epancies may exis
due o AI’s dynamic na u e and luc ua ing socio-poli ical ac o s in luencing b ain d ain.
Addi ionally, inco po a ing di e se quali a i e insigh s, like AI p o essionals’ o policymake s’
expe iences, could ha e en iched ou me hodology.
Fu u e esea ch could del e deepe in o he ollowing a eas: Global Pe spec i e (while ou
s udy ocused on he EU, unde s anding his dynamic in o he egions o on a global scale
could p o ide mo e comp ehensi e insigh s); sec o al Analysis (AI’s in eg a ion a ies ac oss
sec o s. Fu u e s udies could examine how b ain d ain a ec s AI eadiness in speci ic go e n-
men sec o s, such as heal h o de ense).
Au ho con ibu ions
The au ho s (ICI and AS) con ibu ed equally o he elabo a ion o his esea ch.
Disclosu e s a emen
Au ho s ha e no compe ing inancial, p o essional, o pe sonal in e es s om o he pa ies.
Re e ences
Ag awal, A., Gans, J., & Gold a b, A. (2019). The economics o a i icial in elligence: An agenda. Uni e si y
o Chicago P ess. h ps://doi.o g/10.7208/chicago/9780226613475.001.0001
