Liu, Wei; Cao, Mengxiao; Zheng, Jianwen; Zhang, Zuopeng
A icle
Independence o in e dependence: The ole o a i icial
in elligence in co po a e en y mode o o e seas ene gy
in es men s
Jou nal o Inno a ion & Knowledge (JIK)
P o ided in Coope a ion wi h:
Else ie
Sugges ed Ci a ion: Liu, Wei; Cao, Mengxiao; Zheng, Jianwen; Zhang, Zuopeng (2024) : Independence
o in e dependence: The ole o a i icial in elligence in co po a e en y mode o o e seas ene gy
in es men s, Jou nal o Inno a ion & Knowledge (JIK), ISSN 2444-569X, Else ie , Ams e dam, Vol. 9,
Iss. 3, pp. 1-11,
h ps://doi.o g/10.1016/j.jik.2024.100518
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/327421
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-nc-nd/4.0/
Independence o in e dependence: The ole o a ificial in elligence in
co po a e en y mode o o e seas ene gy in es men s
Wei Liu
a
, Mengxiao Cao
a
, Jianwen Zheng
b
, Jus in Zuopeng Zhang
c,
*
a
Business School, Qingdao Uni e si y, Qingdao 266071, PR China
b
Depa men o S a egic Managemen and O ganiza ion, In e na ional Business School Suzhou, Xi’an Jiao ong-Li e pool Uni e si y, Suzhou, PR China
c
Depa men o Managemen , Coggin College o Business, Uni e si y o No h Flo ida, Jackson ille, FL 32224, USA
ARTICLE INFO
A icle His o y:
Recei ed 1 Ap il 2024
Accep ed 17 July 2024
A ailable online 28 July 2024
ABSTRACT
A ificial in elligence (AI) echnology has significan ly ans o med co po a e beha io in he ene gy sec o
by enhancing he capaci y and e ficiency o in o ma ion ansmission and big da a analysis. Howe e , he e
is s ill a limi ed unde s anding o how AI influences new ma ke en y mode s a egies in o e seas ene gy
in es men s. D awing on in o ma ion p ocessing heo y, we p opose ha fi ms wi h ad anced AI echnology
exhibi supe io da a p ocessing capabili ies, which can help ene gy companies mi iga e he unce ain ies
associa ed wi h en e ing a o eign ma ke and encou age hem o choose a wholly-owned en y mode. We
u he hypo hesize ha he s a e owne ship o fi ms, he poli ical a fini y be ween home and hos coun ies,
and he isk p e e ences o fi m execu i es se e as mode a ing ac o s. Using a sample o Chinese-lis ed mul-
ina ional fi ms in he ene gy sec o om 2010 o 2021, ou empi ical esul s s ongly suppo hese p edic-
ions. These findings con ibu e o he eme ging and c ucial li e a u e on he impac o AI echnology on
fi ms’o e seas in es men beha io , pa icula ly in he ene gy sec o .
© 2024 The Au ho (s). Published by Else ie España, S.L.U. on behal o Jou nal o Inno a ion & Knowledge.
This is an open access a icle unde he CC BY-NC-ND license
(h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/)
Keywo ds:
A ificial in elligence
Ene gy in es men s
En y mode
S a e owne ship
Poli ical a fini y
Risk p e e ences
JEL classifica ion:
F23
O33
Q43
In oduc ion
When expanding o e seas, fi ms should conside no only he
loca ion choice bu also he en y mode o in e na ionaliza ion (Hu z-
schen eu e , Ma & Kleindiens , 2020). En y mode is an impo an
s a egic decision in he in e na ional expansion p ocess o mul ina-
ional en e p ises (MNEs) (B ou he s, 2013), which g ea ly influences
fi ms’in e na ionaliza ion beha io and pe o mance (Cui & Jiang,
2009;Tihanyi, G i fi h & Russell, 2005). None heless, he e is al eady
a conside able amoun o li e a u e in es iga ing he d i e s o en y
mode om di e en pe spec i es, such as ins i u ional ac o s (Ang,
Benischke & Doh, 2015;He n
andez & Nie o, 2015), cul u al dis ance
(Chang, Kao, Kuo & Chiu, 2012;Tihanyi e al., 2005), and fi m esou -
ces (Meye , W igh & P u hi, 2009).
The deep in eg a ion o digi aliza ion wi h global business has
also made he digi aliza ion o MNEs a majo end in in e na ional
business s udies (Ghau i, S ange & Cooke, 2021;Luo, 2021). In his
field, he s udy o he in e na ionaliza ion o bo n-digi al fi ms has
a isen wi h a ocus on digi al isks (Kim & Ca usgil, 2020;Luo, 2022;
Monaghan, Tippmann, & Co iello, 2020). Meanwhile, he e ha e
been conce ns abou he use o digi al echnologies in MNEs, such as
he applica ions o digi aliza ion in a eas such as he speed o fi ms’
in e na ionaliza ion (Deng, Zhu, Johanson & Hilme sson, 2022;Lee,
Falaha & Sia, 2019), business models (Reim, Yli-Vii ala, A as uo i &
Pa ida, 2022), me ge and acquisi ion (Wang, Yuan, Huang, Liu &
Zhang, 2024), and global s a egy (Meye , Li & B ou he s, 2023). P e-
ious s udies ha e also analyzed he impac o digi iza ion on fi ms’
choice o en y modes (Ekeledo & Si akuma , 2004b;Henna , 2022;
Li, Zhang, Fan & Li, 2021).
Howe e , cu en in e na ional business schola s ha e p ima ily
ocused on digi iza ion as a collec ion o digi al echnologies, wi h li -
le a en ion paid o he specific echnologies ha cons i u e digi iza-
ion, which may ha e unique implica ions o o ganiza ions (Ahi,
Sinko ics, Shildibeko , Sinko ics & Mehandjie , 2022;Ciulli & Kolk,
2023). Specifically, a ificial in elligence (AI) has he abili y o p ocess
da a, which can help o ganiza ions make be e p edic ions and
explo e po en ial pa e ns (Alnsou , Johnson, Albiz i & Ha ouche,
2023;Bosma & an Wi eloos uijn, 2024;Ha ouche, Quinio &
Bugio i, 2023). Ye , knowledge abou whe he AI a ec s an o ganiza-
ion’s decision o en y mode is limi ed. This is su p ising as AI is
* Co esponding au ho .
E-mail add ess: [email p o ec ed] (J.Z. Zhang).
h ps://doi.o g/10.1016/j.jik.2024.100518
2444-569X/© 2024 The Au ho (s). Published by Else ie España, S.L.U. on behal o Jou nal o Inno a ion & Knowledge. This is an open access a icle unde he CC BY-NC-ND license
(h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/)
Jou nal o Inno a ion & Knowledge 9 (2024) 100518
Jou nal o Inno a ion
&Knowledge
h ps://www.jou nals.else ie .com/jou nal-o -inno a ion-and-knowledge
eme ging as an impo an s a egic esou ce in he in e na ional busi-
ness a ena, ealizing alue o MNEs (Ghau i e al., 2021).
To add ess his esea ch gap, we ocus on he impac o AI echnol-
ogy on he en y mode o MNEs in o o eign ma ke s. Using Chinese
publicly lis ed MNEs in ene gy sec o om 2010 o 2021 as he sam-
ple, ou findings p o ide a su ficien ly s ong confi ma ion o he
a gumen ha AI echnology can acili a e he en y o ene gy fi ms
in o o eign ma ke s in a wholly-owned manne . G ounded in he
in o ma ion p ocessing heo y (Egelho , 1991), we a gue ha ene gy
fi ms ha ing highe le els o AI echnology imply be e in o ma ion
p ocessing capabili ies, which can be e ec i e in an icipa ing and
minimizing unce ain y si ua ions aced by fi ms, especially when
hey en e he hos coun ies. Reduced unce ain y leads o lowe
business isks and inc eased iabili y o fi ms in o eign ma ke s,
esul ing in g ea e confidence and ad an age in choosing he
wholly-owned mode. In addi ion, since he in e na ionaliza ion o
ene gy fi ms will influence he p ofi and s a egic secu i y o he
hos coun ies, we a gue ha he di e en poli ical a fini y and he
co po a e s a e owne ship can mode a e he linea ela ionship
be ween he le el o AI echnology and o eign ma ke en y mode in
di e en ways. Finally, since fi m execu i es can significan ly impac
fi m s a egies, we also p opose he mode a ing ole o he isk p e -
e ences o fi m execu i es.
We ocus on he ene gy sec o in China because China has become
one o he mos impo an global ene gy consume s, bu mo e han
hal o i s ene gy demand mus be supplied om ab oad. In addi ion,
Chinese ene gy fi ms ha e in es ed ex ensi ely ab oad, and he go -
e nmen has es ablished in e na ionalized ene gy coope a ion mech-
anisms. Thus, Chinese ene gy fi ms ha e mani es ed a globalized
layou , and hei o eign ma ke en y mode has gene al heo e ical
and p ac ical implica ions.
Ou s udy con ibu es significan ly o he ela ed li e a u e. Fi s ,
ou in es iga ion ocuses on an impo an s a egy o in e na ionali-
za ion: en y mode. While he e is p io li e a u e on AI echnology
in MNEs, i mainly ocuses on i s impac on MNEs’expo , ma ke ing,
pe o mance, and human esou ce managemen (Denicolai, Zucchella
& Magnani, 2021;Hossain, Agniho i, Rushan, Rahman & Sumi, 2022;
Malik, De Sil a, Budhwa & S ikan h, 2021;Yang, Lee & Chang, 2023).
Second, we in es iga e a c ucial componen o digi iza ion: AI ech-
nology. By del ing in o he AI module o digi iza ion, we gained new
insigh s in o he ole o digi iza ion in de eloping in e na ionaliza-
ion s a egies o MNEs. In addi ion, his s udy expands en y mode
s a egies in o e seas in es men in he ene gy indus y om an
in o ma ion echnology pe spec i e, which specifically ex ends ou
unde s anding o he pa h o “how echnology a ec s fi ms’in e na-
ional en y mode by influencing knowledge p ocessing.”
Theo e ical backg ound
A ificial in elligence and in o ma ion p ocessing
As a ep esen a i e o he new gene a ion o in o ma ion echnol-
ogy, AI is ecei ing ex ensi e a en ion om schola s and p ac i-
ione s (Bosma & an Wi eloos uijn, 2024;Collins, Dennehy, Conboy
& Mikale , 2021). AI is desc ibed as he use o heo ies, echniques,
echnologies, and applica ion sys ems o eplica e, enhance, and aug-
men human in elligence (Rai, Cons an inides & Sa ke , 2019). Cu -
en ly, MNEs a e using AI mo e equen ly o manage hei
ope a ions and p o ide po en ial solu ions (Luo & Zah a, 2023).
Specifically, AI echnology can p edic changes in in e na ional
ade based on he in o ma ion a ailable, hus helping fi ms make
be e decisions (Ra hje, Ka ila & Reineke, 2024;Yang e al., 2023).
Machine lea ning skills in AI echnology ha e aided banking o gani-
za ions in de ec ing c edi ca d aud in he global financial sec o
(Chinn, Kaplan & Weinbe g, 2014). In addi ion, AI echnology can
la gely imp o e a fi m’s o e seas sales pe o mance. Acco ding o
Denicolai e al. (2021)), AI echnology can enhance fi ms’unde s and-
ing o o eign ma ke s by u ilizing da a analy ics o suppo hei
in e na ional ma ke ing campaigns as well as emo e con ac s wi h
local clien s. In he global managemen o MNEs, human esou ce
managemen (HRM) is a c i ical unc ion (Cooke, Wu, Zhou, Zhong &
Wang, 2018;Schule , Dowling & De Cie i, 1993). AI echnology in
HRM boos s in o ma ion flow and exchange, inc eases he e u n on
in es men in HRM, and p o ides a cus omized expe ience o
employees (Malik e al., 2021;Malik, Budhwa , Pa el & S ikan h,
2022;V on is e al., 2022).
Acco ding o in o ma ion p ocessing heo y (Egelho , 1991), fi ms
can be hough o as in o ma ion p ocessing sys ems wi h unce -
ain y wo k. In his si ua ion, AI echnology is c ucial o add ess he
unce ain y (Da , 1992). On he one hand, AI echnology is capable o
iden i ying and analyzing isks on he basis o o ganiza ions p ocess-
ing in o ma ion and p o iding o ganiza ions wi h decisions (conclu-
sions based on algo i hmic delibe a ion o a ailable da a) and
solu ions (al e na i e cou ses o ac ion o sol e issues) (Flasi
nski,
2016). On he o he hand, AI echnology is capable o analyzing and
p edic ing known da a and in o ma ion and imp o ing he o ganiza-
ion’s in o ma ion p ocessing capabili ies o cope wi h in e nal com-
plexi y and en i onmen al unce ain y.
En y mode selec ion in esponse o new ma ke unce ain y
Modes o en y in o o eign ma ke s, which we call he en y
mode in his s udy, seem o be he mos popula di ec ions o schol-
a s o esea ch in e na ional business (We ne , 2002). En y mode is
conside ed “a s uc u al ag eemen ha allows a fi m o implemen
i s p oduc ma ke s a egy in a hos coun y ei he by ca ying ou
only he ma ke ing ope a ions (i.e., ia expo modes), o bo h p o-
duc ion and ma ke ing ope a ions he e by i sel o in pa ne ship
wi h o he s (con ac ual modes, join en u es, wholly owned ope a-
ions)”(Sha ma & E amilli, 2004). En y mode can be ca ego ized as
equi y and non-equi y, and hese wo ypes di e significan ly in
e ms o in es men equi emen s and con ol. Compa ed o non-
equi y mode (e.g., con ac ual modes such as licenses), equi y mode
(e.g., join en u es and wholly-owned en u es) means ha fi ms
need o make g ea e in es men commi men s and gain a highe
con ol le el in he o e seas business (Pan & Tse, 2000). This s udy
ocuses on he di e en en y modes be ween wholly owned
subsidia ies and join en u es. Bo h en y mode choices ep esen
di e en isk- aking, commi men s, and e u ns. Wholly owned
subsidia ies enable mul ina ional en e p ises o main ain ull owne -
ship and con ol o e hei in e na ional ope a ions, while he join -
en u e ype means ha he fi ms need o coope a e wi h local fi ms
and sha e he equi y (B ou he s & Henna , 2007;Slangen & Van
Tulde , 2009). Since coope a ion wi h local fi ms can ake ad an age
o he local influence o he pa ne fi m and i s s eng hs in ins i u-
ional legi imacy and in o ma ion ad an ages, he unce ain y and
isk aced by he join en u e will be educed, and co espondingly,
he e u ns ecei ed by he pa en fi m will also be educed (Kim &
Hwang, 1992).
When MNEs es ablish subsidia ies in a o eign coun y, hey will
ace unknown si ua ions and en i onmen al unce ain y. F om he
ansac ion cos heo y pe spec i e, he e a e wo di e en unce -
ain ies. In e nal unce ain y usually comes om he employees’pe -
o mance, while ex e nal unce ain y is always caused by he
unp edic abili y o he hos en i onmen , such as he di e en poli-
cies o special laws (B ou he s & Henna , 2007). Schola s ha e dem-
ons a ed ha when o eign ma ke s ha e high le els o unce ain y,
pa en fi ms should a o wholly-owned subsidia ies a he han
join en u es. The high le el o equi y can help subsidia ies p o ec
p op ie a y ad an ages when hey fi s build up in a new coun y,
and hen, hey can communica e di ec ly o educe ansac ion cos s.
The e o e, fi ms can o se ex e nal unce ain y h ough in eg a ion
W. Liu, M. Cao, J. Zheng e al. Jou nal o Inno a ion & Knowledge 9 (2024) 100518
2
decisions (B ou he s, 2013). Mo eo e , he di e en ins i u ional
en i onmen s a e ano he impo an unce ain y ac o influencing
co po a e s a egic decision-making in new ma ke s, especially he
en y mode s a egy. Acco ding o he ins i u ional heo y, fi ms
mus comply wi h he special ules o gain legi imacy om he local
go e nmen (Gao & Ha si, 2015). In o he wo ds, a s able and eason-
able ins i u ional en i onmen means a good business en i onmen ,
which will b ing less poli ical and legal unce ain y. Thus, MNEs will
choose he join en u e mode o es ablish he subsidia ies because
o he highe pe o mance and he less isk.
MNEs expanding ab oad mus con end wi h a ange o unknowns
in he new ma ke ha limi hei ope a ions in he home ma ke
(Hill, 2008). Compa ed o domes ic i als, new o eign en an s ace
he “liabili y o o eignness”(LoF) in he hos na ion. This d awback
esul s om he MNE’s inexpe ience wi h he ins i u ions, en i on-
men , and cul u e o he hos ma ke as well as om highe cos s and
disc imina ion (Eden & Mille , 2004). Fo MNEs, o e coming he LoF
is an indispensable condi ion o fi ms o in es di ec ly ab oad (Ba -
kema & D ogendijk, 2007). The e o e, when en e ing new ma ke s,
fi ms need an op imal ma ke en y mode ha can help hem educe
unce ain y in un amilia en i onmen s and be able o ully u ilize
hei supe io capabili ies (Chen, 2006).
Hypo heses de elopmen
A ificial in elligence and en y mode selec ion
As fi ms expand globally, leade s o MNEs ace he challenge o
managing complex ope a ions ac oss di e en egions and coun ies
(Eden & Nielsen, 2020). MNEs ha e mo e complex in e nal sys ems
and coo dina ion wi h ex e nal s akeholde s han domes ic fi ms
(Beni o, Pe e sen & Welch, 2019). Meanwhile, MNEs may ace in i-
ca e and un amilia ex e nal ac o s when conduc ing business
ab oad (Bha dwaj, Die z & Beamish, 2007). Fo example, he di e si y
o o eign business en i onmen s, as well as he dynamic and un a-
milia na u e o ma ke en i onmen s, inc ease he unce ain y aced
by fi ms, equi ing fi ms o be able o analyze and p ocess la ge
amoun s o in o ma ion in a imely manne and o imp o e hei
in o ma ion p ocessing capabili ies in o de o suppo hei deci-
sion-making p ocess (Egelho , 1991;Kano, Tsang & Yeung, 2020).
AI is an eme ging and highly e ec i e echnology o in o ma ion
and da a p ocessing (Bosma & an Wi eloos uijn, 2024). I ecog-
nizes, lea ns, and p ocesses ex e nal and in e nal in o ma ion gene -
a ed by o ganiza ions, helping fi m manage s make in o med
decisions and ake app op ia e ac ions (Ahmad e al., 2021). On one
hand, AI echnology can help fi ms p ocess in o ma ion mo e e fi-
cien ly when en e ing o eign ma ke s. This can aid in quickly unde -
s anding new ma ke s, analyzing and p edic ing u u e ma ke
ends, and educing he cos s associa ed wi h amilia izing hem-
sel es wi h he en i onmen . On he o he hand, i can also ge ahead
o compe i o s and gain an ad an age o e i s domes ic coun e pa s
when dealing wi h changing en i onmen al condi ions. The e o e, AI
echnology can assis fi ms in educing en i onmen al unce ain y,
o e coming he LoF when en e ing he hos ma ke s, and enhancing
hei iabili y. In his case, fi ms a e mo e likely o choose a wholly
owned model o en y in o de o maximize he e u n on in es men
wi hin he abili y o bea ce ain isks. We he e o e p opose:
Hypo hesis 1. Fi ms wi h high le els o AI echnology end o use
wholly-owned en y mode o he o eign ma ke s, while fi ms wi h low
le els o AI echnology end o use join en u e en y mode.
S a e owne ship as he mode a o
The ea u es o he ins i u ional se ing ha e a significan impac
on he ene gy sec o and he linked businesses. Ene gy is a c i ical
esou ce o he coun y and is o en influenced by na ional policies
and poli ical objec i es, as no ed by F ynas and Paulo (2007). S a e-
owned en e p ises (SOEs) may ha e di e en objec i es han non-
SOEs and may be seen as ep esen a i es o hei home go e nmen s.
S a e owne ship may lead hese fi ms o in es ab oad mo e o poli -
ical easons han o economic gain Cue o-Cazu a, Inkpen, Musac-
chio and Ramaswamy (2014). Meanwhile, China’s pu sui o o eign
ene gy sou ces has aised se ious conce ns in o he coun ies. Fo
ins ance, CNOOC (a Chinese s a e-owned fi m) a emp ed he acquisi-
ion o Unocal (a US oil fi m) in 2005, igge ing significan poli ical
opposi ion in he Uni ed S a es, as he ene gy indus y is a sensi i e
sec o ha has come unde e en mo e sc u iny, which he US go e n-
men belie es could unde mine US na ional secu i y and hu i al
ene gy supplies (Wan & Wong, 2009). As a esul , s a e owne ship
may igge legi imacy conce ns o he hos go e nmen and o he
s akeholde s, aising poli ical, na ional secu i y, and economic con-
ce ns, as well as suspicions. This inc eases he poli ical and social
unce ain y ha fi ms ace in hos coun ies and unde mines he
compe i i e ad an age ha AI c ea es o fi ms in hos coun y ma -
ke s.
While pu suing economic in e es s, SOEs also need o conside
o he non-economic ac o s, such as na ional s a egic objec i es and
indus ial policies (Cue o-Cazu a e al., 2014). This may lead SOEs
o weigh o he non-economic ac o s in he decision-making p ocess
o he en y mode in addi ion o he ma ke ac o s analyzed by AI.
The influence o AI echnology in he decision-making o SOEs is less
weigh ed compa ed o ha o non-SOEs, which educes he impac o
AI echnology on fi ms’choice o he wholly owned en y mode.
In addi ion, SOEs may ace mo e cons ain s and complex deci-
sion-making p ocesses (Budiman, Lin & Singham, 2009;U oyo, Ma i-
min & Mu dano o, 2019), which may educe hei abili y o u ilize AI
echnology o conduc ma ke analysis and o espond apidly o
ma ke changes. In con as , non-SOEs ypically ha e mo e flexible
decision-making mechanisms and can make as e decisions based
on AI analyses. As a esul , he s a e-owned na u e o fi ms may make
hem mo e cau ious and educe hei p e e ence o he wholly-
owned en y mode. We he e o e p opose:
Hypo hesis 2. S a e owne ship weakens he posi i e ela ionship
be ween fi ms’AI echnology and hei wholly-owned en y mode o o -
eign ma ke s.
Poli ical a fini y as he mode a o
When selec ing an en y mode, fi ms mus conside he pa icula
cha ac e is ics o he hos coun y (Shama, 1995;Tihanyi e al.,
2005). In his con ex , fi ms should unde s and he hos coun y’s
ins i u ional en i onmen and examine in e na ional ag eemen s on
bila e al poli ics, including poli ical a fini y conside a ions (Dixon &
Moon, 1993;Li & Vashchilko, 2010).
A highe le el o poli ical a fini y means ha he fi m comes om
a coun y ha sha es mo e o he same na ional in e es s, which will
lead o mo e suppo om he hos coun y’s s akeholde s, and he
fi m’s beha io in he hos coun y is aken “ o g an ed.”As Suchman
(1995) poin s ou , “ aken o g an ed”is, wi hou a doub , he mos
subdued and po en sou ce o legi imacy ound. Fi m beha io ha is
aken o g an ed is a ely subjec o ex e nal in e en ion. Duanmu
(2014) also shows ha coun ies wi h close poli ical ies will gain
mo e us h ough pas in e ac ions in he bila e al con ex in which
FDI beha io akes place. This will p omo e dialog and suppo
be ween home and hos go e nmen s o mi iga e po en ial exp op i-
a ion isks in he hos coun y. When poli ical a fini y is low, he fi m
comes om a coun y wi h mo e di e se na ional in e es s, a which
poin hos coun y s akeholde s will ha e mo e significan secu i y
conce ns abou he ene gy fi m. Fi ms’access o in o ma ion om
he hos coun y may be hinde ed by s akeholde s such as he hos
W. Liu, M. Cao, J. Zheng e al. Jou nal o Inno a ion & Knowledge 9 (2024) 100518
3
go e nmen (Ma io i & Pisci ello, 1995). This may inc ease he cos
and di ficul y o fi ms in ob aining in o ma ion abou he hos coun-
y’s ma ke . On he con a y, highe poli ical a fini y implies a con-
e gence o na ional in e es s and hus educes he legi imacy
p oblems aced by MNEs in he hos coun y (Hasija, Liou & Ells and,
2020). When a high le el o us and goodwill exis s be ween coun-
ies, fi ms ace less p essu e om egula o y sys ems, such as go -
e nmen s, which educes he cos and di ficul y o ob aining
in o ma ion, and fi ms a e likely o ha e access o iche in o ma ion
han in coun ies wi h lowe poli ical a fini y.
The analysis and p edic ion unc ion o AI is based on he p ocess-
ing o his o ical da a, and he mo e da a he e is, he mo e accu a e
he esul s will be. Thus, a high le el o poli ical a fini y be ween he
wo coun ies enables fi ms o ob ain mo e in o ma ion quickly and
easily, allowing AI echnology o ully u ilize i s benefi s while s udy-
ing he hos na ion’s ma ke condi ions. This can help fi ms quickly
unde s and he ma ke in he hos na ion, educe unce ain y, and
es ablish compe i i e ad an ages in he ma ke .
The e o e, fi ms en e ing coun ies wi h high le els o poli ical
a fini y will be e enjoy a ious ad an ages and isk a oidance
b ough by AI echnology and hus will be mo e inclined o selec he
wholly-owned en y mode. We hypo hesize ha :
Hypo hesis 3. Poli ical a fini y s eng hens he posi i e ela ionship
be ween fi ms’AI echnology and hei wholly-owned en y mode o he
o eign ma ke s.
Execu i e isk p e e ence as he mode a o
In co po a e go e nance, execu i es, as he manage s o fi ms,
o en play majo oles in s a egic decision-making. Howe e , acco d-
ing o Chiles and McMackin (1996)) and B ou he s and B ou he s
(2003), manage s migh no be isk-neu al, and isk-a e se manag-
e s migh decide di e en ly om isk-seeke manage s. Wi h he
de elopmen o AI echnology, co po a e decision-make s a e g adu-
ally no icing i as a powe ul and ad an ageous in o ma ion p ocess-
ing ool o decision-making and s a egy o mula ion (Huo &
Chaudh y, 2021;Ra hje e al., 2024). When co po a e execu i es
ha e a highe isk appe i e, hey a e mo e inclined o seek highe
e u ns, meaning hey will ake highe isks associa ed wi h unce -
ain ies in he hos ma ke . Howe e , AI echnology can analyze la ge
amoun s o his o ical da a and p o ide in-dep h ma ke insigh s and
o ecas s, which can help execu i es unde s and he ma ke condi-
ions in he hos coun y and iden i y ma ke oppo uni ies and isks,
hus suppo ing execu i es in de eloping hei s a egies and inc eas-
ing he chances o he fi m’s success in he hos coun y.
When he execu i es’accep ance o isk is high, he in o ma ion
suppo p o ided by AI echnology educes he unce ain y in he
hos ma ke . This enhances he in es men confidence o isk-a e se
execu i es who a e mo e willing o ake ad an age o AI echnology
o pu sue highe e u ns and g ea e ma ke oppo uni ies. Thus,
based on he aluable in o ma ion p o ided by AI echnology, isk-
a e se execu i es a e mo e inclined o selec he wholly-owned en y
mode in he decision-making p ocess. This is because his decision
means fi ms can implemen hei s a egies flexibly and p ofi ably.
The e o e, we hypo hesize ha :
Hypo hesis 4. The execu i e isk p e e ence s eng hens he posi i e
ela ionship be ween fi ms’AI echnology and hei wholly-owned en y
mode o o eign ma ke s.
Me hodology
Sample and da a collec ion
We cons uc ed a da ase o MNEs in China’s ene gy sec o om
2010 o 2021 o es ou hypo heses. China is one o he mos
impo an global ene gy consume s, elying on impo s o mo e
han hal o i s ene gy needs, and hus, Chinese ene gy fi ms a e e y
ac i e in ou wa d o eign di ec in es men (OFDI). A sample o Chi-
nese fi ms in he ene gy sec o lis ed on bo h he Shenzhen s ock
exchanges and Shanghai s ock exchanges was selec ed o analysis.
This is due o he ac ha lis ed fi ms, which a e obliga ed by law o
gi e co ec in o ma ion in hei annual epo s, ha e mo e depend-
able da a on hei ope a ions ab oad and su ficien esou ces o OFDI
(Xia, Ma, Lu & Yiu, 2014). The e o e, he Chinese-lis ed fi m in he
ene gy sec o p o ides an ideal se ing in ou s udy.
We collec ed da a ela ed o all pa en fi ms and o eign subsidia -
ies om he China S ock Ma ke & Accoun ing Resea ch Da abase,
which is a majo sou ce o da a ela ed o he Chinese lis ed fi ms
(Wang & Qian, 2011). In his da abase, da a on AI began o be a ail-
able in 2010, so his yea was he s a ing poin o he sample. We
ocused on he o eign ma ke en y mode o subsidia ies in his
s udy, so we cons uc ed subsidia y-yea da a as he basic indepen-
den obse a ions, and he final sample con ains a o al o 654 o eign
subsidia ies.
Measu emen
Dependen a iable
En y mode. Following p e ious s udies (Bloms e mo, Deo Sha ma
& Sallis, 2006;Ekeledo & Si akuma , 2004b), we used he sha e o
equi y in subsidia ies con olled by he pa en fi m as a measu e o
en y mode. We defined wholly-owned fi ms as hose wi h co po a e
owne ship sha es abo e 95 pe cen equi y and 10−95 pe cen own-
e ship sha es as join en u es. Owne ship sha es below 10 pe cen
a e no included because hese kinds o fi ms can be in e p e ed as
po olios a he han di ec in es men s (Puck, H€
odl, Fila o che ,
Wol & Bade , 2016). Specifically, we measu ed his a iable as he
dummy a iable, aluing 1 o he en y mode o a wholly-owned
subsidia y o en e a o eign ma ke and 0 o he wise.
Independen a iable
AI echnology was used as he independen a iable. We ollowed
he p e ious s udies on fi m’s AI and used he numbe o wo d e-
quencies o keywo ds ela ed o AI in he annual epo s o lis ed
fi ms as a p oxy o he le el o he fi m’s AI echnology (Li e al.,
2023a). We used he equency o AI- ela ed wo ds o lis ed fi ms
om he CSMAR da abase. The keywo d collec ion was cons uc ed
om he pe spec i es o bo h AI echnology and AI echnology in a-
s uc u e o o m a dic iona y o keywo ds ha can eflec he le el
o AI, wi h a o al o 36 keywo ds. Table 1 ep esen s he keywo ds
dic iona y used in he measu emen . These da a a e ob ained om
annual epo s using ex mining me hods, as annual epo s eflec
an o ganiza ion’s cu en s a e o business and co po a e s a egy.
(Klop chenko e al., 2004).
Mode a ing a iables
S a e owne ship. Due o he di e ences in he choice o en y mode
be ween SOEs and non-SOEs (G øgaa d, Rygh & Beni o, 2019), we ol-
lowed he s udy o Duanmu (2014) and used he ac ual p opo ion o
s a e-owned sha es o he fi m o measu e he deg ee o s a e owne -
ship o he pa en fi ms.
Poli ical a fini y. Poli ical ela ions be ween coun ies can influ-
ence fi ms’s a egic choices. We ollowed (Be and, Be schinge and
Se les (2016) and Ga zke (1998) and measu ed i using he consis-
en o ing beha io be ween he wo coun ies in he Uni ed Na ions
Gene al Assembly. We use a con inuous a iable alued be ween 0
and 1, a alue o 1 ep esen s comple ely consis en beha io o o -
ing be ween he fi m’s home coun y and i s a ge hos coun y, and
0 ep esen ing opposi e beha io o o ing (Be and, Be schinge &
Se les, 2016;Fiebe g, Lopa a, Tammen & Tideman, 2021).
W. Liu, M. Cao, J. Zheng e al. Jou nal o Inno a ion & Knowledge 9 (2024) 100518
4
Risk p e e ences o execu i es. Due o he specifici y o he
Chinese capi al ma ke , he p opo ion o isky asse s in indi id-
ualasse sislimi edby heda aacquisi ionand hesmallnum-
be o Chinese execu i es’sha eholdings, which is unsui able o
he Chinese con ex . We ollowed he me hod o Abdel-Khalik
(2007) and Zhao, Niu and Chen (2022) o measu e his a iable
by calcula ing a comp ehensi e index wi h mul iple indica o s.
The a io o isky asse s o o al asse s, gea ing a io, co e p ofi -
abili y a io, e ained ea ning a e, sel - unding sa is ac ion a e,
and capi al expendi u e a e a e he six indica o s we chose o
his s udy o assess execu i es’ isk appe i e. These indica o s
a e d awn om he fi e aspec s o asse s uc u e, sol ency,
p ofi s uc u e, p ofi dis ibu ion, and cash flow. Then, we cal-
cula ed hem using he p incipal componen analysis me hod.
The posi i e and la ge alues o execu i e isk appe i e mean
ha execu i es a e mo e inclined o ake high isks, and nega-
i e and smalle alues mean ha execu i es a e mo e isk-
a e se.
Con ol a iables
We also e e ed o a ange o s udies on en y mode, including
di e en le els o a iables o con ol o po en ial con ounding
e ec s (He n
andez & Nie o, 2015;Kao & Kuo, 2017).
A he fi m le el, we fi s included O e seas expe ience, which is
measu ed as he dummy a iable, coded 1 i he execu i e eam
membe s ha e o e seas s udy o enu e expe ience and 0 o he wise.
P io esea ch has demons a ed ha fi ms wi h execu i e eams
possessing o e seas expe ience a e inclined o op o wholly-owned
subsidia ies (Nielsen & Nielsen, 2011). Subsequen ly, we accoun ed
o fi m size, which was quan ified by he loga i hm o a fi m’s annual
o al asse s. Fi m age is calcula ed by he numbe o yea s since i s
es ablishmen . La ge fi ms wi h mo e ex ensi e expe ience end o
possess g ea e knowledge o in e na ional ope a ions and supe io
managemen capabili ies (Jain, Thuk al, & Paul, 2024). Deb a io was
u ilized as a me ic, exp essed as he pe cen age o o al liabili ies o
o al asse s, o mi iga e he impac o capi al s uc u e on he deci-
sion-making o en y modes. We also used ROA o measu e Financial
pe o mance.
A he coun y le el, we included Tu no e , measu ed as he
dummy a iable, coded as 1 i he hos coun y leade ship u no e
occu s in he yea o en y in o he hos coun y and 0 o he wise.
Ins i u ional dis ance has always been a majo conce n o in e na-
ional business schola s in s udying he an eceden s o en y mode
(Ang e al., 2015;B ou he s, 2013). We measu ed i as he Wo ld
Bank’s Global Go e nance Index (Kogu and Singh (1988).Economic
dis ance can influence he en y mode o MNEs (Tao, Zhanming &
Xiaoguang, 2013), and we measu ed i by he loga i hms o GDP pe
capi a o he home coun y o each hos coun y.
Es ima ion me hod
We u ilized a logis ic eg ession model o examine he impac o
AI echnology on he likelihood ha MNEs will choose a wholly
owned subsidia y o join en u e, gi en ha he dependen a iable
(En y mode) is a dummy a iable (L
opez-Dua e & Vidal-Su
a ez,
2013). Specifically, he eg ession model was es ima ed as ollows:
En ymode ¼b0þb1AI echnology þb2S a eowne ship þb3Poli icala ini y
þb4Riskp e e enceso execu i es þb5AI echnology
S a eowne ship þb6AI echnology Poli icala ini y
þb7AI echnology Riskp e e enceso execu i es þb8Con ols þe
whe e b1was used o es Hypo hesis 1, and b5,b6, and b7we e used
o es Hypo heses 2−4 sepa a ely.
Resul s
Desc ip i e s a is ics and co ela ion esul s
Table 2 shows he in o ma ion on he desc ip i e s a is ics and
co ela ion esul s o all a iables. The coe ficien o he co ela ion
be ween AI echnology and en y mode is 0.09, wi h significance a
he 1 pe cen le el. The majo i y o he co ela ion coe ficien s
be ween any wo a iables all below 0.5. Addi ionally, upon assess-
ing he a iance infla ion ac o o he models, we obse ed ha bo h
he highes and mean alues emained unde 5. All he abo e esul s
indica e ha he po en ial mul icollinea i y issue does no exis in
ou s udy.
Hypo heses es ing
The findings pe aining o ou ini ial hypo hesis a e exhibi ed in
Table 3. Model 1 se es as he ounda ional model inco po a ing all
pe inen con ol a iables. In Model 2, we find ha he coe ficien o
AI echnology is s a is ically posi i e wi h significance a he 1 pe cen
le el (b¼0:46;p<0:01). This confi ms ou hypo hesis ha fi ms’AI
echnology will imp o e hei abili y o p ocess in o ma ion. The e-
o e, fi ms do no need o coope a e wi h local fi ms o educe isk,
which is consis en wi h he iew ha fi ms end o choose wholly-
owned fi ms when hey ace less unce ain y.
In Model 3, ou examina ion seeks o de e mine he nega i e
mode a ing e ec o s a e owne ship. Ou analysis e eals ha he
coe ficien o he in e ac ion e m be ween AI echnology and S a e
owne ship is nega i e, wi h he significance a he 1 pe cen le el
(b¼0:02;p<0:01). This confi ms ha s a e owne ship can nega-
i ely mode a e he main ela ionship, and hypo hesis 2 can be
p o ed. The finding indica es ha as he le el o s a e owne ship in
an ene gy fi m in ensifies, he le el o egula o y and legi imacy con-
ce ns in he hos coun y amplifies, he eby diminishing he a o able
impac o AI echnology.
In Model 4, we es whe he s ong poli ical a fini y posi i ely
mode a es he ole o fi ms’AI echnology in hei wholly-owned
en y mode. The findings e eal ha he coe ficien o he in e ac ion
e m be ween AI echnology and Poli ical a fini y is posi i e wi h he
significance a he 1% le el (b¼0:96;p<0:01). In coun ies wi h
high poli ical a fini y, specifically close poli ical ela ions wi h China,
he s onge he ole played by ene gy fi ms’AI echnology in choos-
ing wholly-owned en y mode.
Table 1
Keywo ds o he a ificial in elligence.
Types Keywo ds
A ificial in elligence echnology A ificial In elligence, Business In elligence, Image Unde s anding, In es men Decision Aids, In elligen Da a Analy ics, In elligen
Robo s, Machine Lea ning, Deep Lea ning, Seman ic Sea ch, Biome ics, Face Recogni ion, Speech Recogni ion, Iden i y Ve ifica ion,
Au onomous D i ing, Na u al Language P ocessing, Supe ised Lea ning, Machine T ansla ion, OCR Technology, Compu e Vision,
Machine Vision, Robo ics, In elligen Ques ion and Answe , Expe Sys ems, Neu al Ne wo ks, Lea ning Algo i hms, Au oma ed Rea-
soning, D ones
A ificial In elligence In as uc u e A ificial In elligence Labs, A ificial In elligence Pla o ms, A ificial In elligence Facili ies, A ificial In elligence De ices, A ificial In elli-
gence In as uc u e, Robo s, A ificial In elligence Sys ems, In elligen Te minals, In elligen In o ma ion Sys ems
W. Liu, M. Cao, J. Zheng e al. Jou nal o Inno a ion & Knowledge 9 (2024) 100518
5
Finally, in Model 5, we es he posi i e mode a ing e ec o exec-
u i e isk p e e ences. The coe ficien o he in e ac ion e m
be ween AI echnology and he Risk p e e ences o execu i es is posi-
i e, wi h he significance a he 1 pe cen le el (b¼0:24;p<0:01).
This indica es ha he highe he execu i es’ isk ole ance, he
mo e inclined hey a e owa d wholly subsidia y en y s a egies.
The pu sui o isk makes execu i es mo e likely o choose a high-
commi men en y mode, and on he o he hand, isk-a e se execu-
i es may inc ease hei confidence in AI echnology. The e o e,
hypo hesis 4 can be suppo ed.
Robus ness check
Ins ead o he logis ic eg ession model, we used an al e na i e
model o es he obus ness o he esul s. We used he P obi eg es-
sion model as he al e na i e model, as i is also es ima ed by he
dummy a iable as he dependen a iable in he model. Table 4
epo s hese obus ness esul s. The coe ficien o AI echnology in
Model 1 is posi i e wi h he significance a he 1 pe cen le el, and
he coe ficien s o he in e ac ion e ms, including AI echnology wi h
S a e owne ship in Model 2, Poli ical a fini y in Model 3, and Risk
Table 2
Desc ip i e s a is ics and co ela ion ma ix.
Va iables Mean S. D. (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13)
(1) En y mode 0.70 0.46 1.00
(2) AI echnology 0.67 2.74 0.09 1.00
(3) S a e owne ship 4.34 13.62 0.04 0.05 1.00
(4) Poli ical a fini y 0.67 0.20 0.02 0.06 0.02 1.00
(5) Risk p e e ences o execu i es 0.14 0.52 0.05 0.24 0.03 0.04 1.00
(6) O e seas expe ience 0.53 0.50 0.18 0.01 0.01 0.15 0.19 1.00
(7) Fi m size 23.85 1.61 0.02 0.14 0.14 0.48 0.03 0.11 1.00
(8) Fi m age 16.83 6.07 0.08 0.02 0.03 0.41 0.11 0.18 0.49 1.00
(9) Deb a io 0.40 0.17 0.01 0.14 0.02 0.49 0.06 0.02 0.44 0.09 1.00
(10) Financial pe o mance 0.02 0.05 0.02 0.09 0.12 0.00 0.13 0.07 0.28 0.18 0.24 1.00
(11) Tu no e 0.22 0.42 0.02 0.01 0.08 0.02 0.11 0.03 0.22 0.12 0.05 0.10 1.00
(12) Ins i u ional dis ance 2.69 1.54 0.08 0.12 0.49 0.15 0.04 0.04 0.22 0.12 0.11 0.06 0.10 1.00
(13) Economic dis ance 4.59 0.68 0.14 0.03 0.38 0.04 0.11 0.01 0.18 0.26 0.09 0.04 0.00 0.19 1.00
No e: N= 654; Co ela ions g ea e han |0.01| a e significan a he 0.05 le el.
Table 3
Main esul s.
Va iables Model 1 Model 2 Model 3 Model 4 Model 5
S a e owne ship 0.01 0.01 0.15*** 0.02* 0.01
(0.01) (0.01) (0.05) (0.01) (0.01)
Poli ical a fini y 0.92 1.30* 1.20* 8.11*** 1.32*
(0.63) (0.70) (0.70) (2.31) (0.70)
Risk p e e ences o execu i es 0.60** 0.56* 0.52* 0.52* 1.12
(0.29) (0.31) (0.31) (0.31) (0.68)
O e seas expe ience 1.00*** 1.30*** 1.35*** 1.37*** 1.31***
(0.23) (0.26) (0.26) (0.26) (0.26)
Fi m size 0.28*** 0.40*** 0.42*** 0.43*** 0.39***
(0.09) (0.10) (0.10) (0.10) (0.10)
Fi m age 0.05** 0.08*** 0.09*** 0.08*** 0.08***
(0.02) (0.03) (0.03) (0.03) (0.03)
Deb a io 1.67** 2.30*** 2.25*** 2.28*** 2.32***
(0.76) (0.81) (0.81) (0.81) (0.81)
Financial pe o mance 1.73 2.89 2.93 1.73 2.86
(1.99) (2.14) (2.15) (2.14) (2.15)
Tu no e 0.20 0.03 0.01 0.01 0.01
(0.23) (0.23) (0.23) (0.23) (0.23)
Ins i u ional dis ance 0.26*** 0.33*** 0.33*** 0.34*** 0.33***
(0.08) (0.09) (0.08) (0.08) (0.09)
Economic dis ance 0.43** 0.52*** 0.54*** 0.57*** 0.53***
(0.19) (0.20) (0.21) (0.21) (0.20)
AI echnology 0.46*** 0.44*** 0.49*** 0.45***
(0.11) (0.11) (0.11) (0.10)
AI le el £S a e owne ship 0.02***
(0.01)
AI le el £Poli ical a fini y 0.96***
(0.31)
AI le el £Risk p e e ences o execu i es 0.24***
(0.09)
Yea dummies Yes Yes Yes Yes Yes
Cons an 0.99 2.10 3.08 1.80 1.57
(1.93) (2.00) (2.07) (2.16) (1.98)
Pseudo R
2
0.10 0.13 0.14 0.14 0.13
Obse a ions 654 654 654 654 654
No e: S anda d e o s a e in pa en heses; *** p<0.01, ** p<0.05, * p<0.1.
W. Liu, M. Cao, J. Zheng e al. Jou nal o Inno a ion & Knowledge 9 (2024) 100518
6
p e e ences o execu i es in Model 4 a e s a is ically significan and
align wi h he p ima y esul s p esen ed in Table 3. The e o e, hese
esul s a e s able when using he al e na i e model.
In addi ion, we used an al e na i e measu e o ou impo an a i-
able, AI echnology, o check whe he he esul s we e obus . In con-
as o he main measu emen using ex analy ics, we ollowed he
me hods o Acemoglu and Res epo (2018) and hen measu ed he
ac ual in es men o AI in he fi ms by cap u ing he use o indus ial
obo s. Specifically, he a iable was calcula ed as he a io o he
book alue o indus ial obo s o o al employees. Model 5 o Table 4
epo s he esul s ha he coe ficien o AI echnology is posi i e
wi h he significance a he 1% le el, suppo ing he obus ness o ou
main esul s.
Endogenei y es
Po en ial endogenei y issues, specifically e e se causali y
be ween AI echnology and en y mode, may in e e e wi h ou
esul s and should be add essed wi h mo e es s. We adop ed he
P obi model wi h con inuous endogenous co a ia es (IV-P obi ) o
elimina e he e ec o endogenei y. We ca e ully selec ed wo ins u-
men a iables: Densi y o fibe op ic cables and Densi y o mobile base
s a ions in he p o ince whe e he pa en fi ms a e loca ed (Du, Hou,
Zhou & Ren, 2022;Hossain e al., 2022;Wang, Chen & Chen, 2024).
These wo a iables can eflec he le el o in o ma ion echnology
in as uc u e in he egion and a e highly co ela ed wi h AI echnol-
ogy while unlikely o a ec fi ms’o e seas en y mode di ec ly.
Table 5 epo s he esul s o he endogenei y es . Amemiya-Lee-
Newey es in he o e -iden ifica ion es does no ejec he o iginal
hypo hesis (p¼0:36), indica ing he exogenous ins umen s we used
in ou s udy. The esul s o he Wald es (p¼0:000) sugges ha he
ins umen al a iables we used a e no weak ins umen al a iables.
Model 1 epo s he eg ession esul s o he IV-P obi fi s s age,
showing ha he ins umen al a iables a e significan ly associa ed
wi h AI echnology. Models 2−5 exhibi he second-s age eg ession
esul s. The coe ficien o AI echnology is s a is ically posi i e and sig-
nifican in Model 2 (b¼0:33;p<0:001), sugges ing ha ou main
ela ionship emains he same a e add essing he po en ial endoge-
nei y issues. The esul s o Models 3−5 sugges ha all mode a ing
esul s a e s ill obus excep o s a e owne ship.
Discussion
Wi h he wa e o digi al globaliza ion, fi ms in eme ging ma ke s
a e ac i ely seeking global expansion. I is impo an o u he
unde s and he impac o digi iza ion on business s a egy decisions
(Menz e al., 2021). P e ious s udies ha e ypically ocused on he
impac o digi iza ion as a whole on he business as a uni o on spe-
cific aspec s o i (Lee e al., 2019;To o a, Chie ici, B iamon e & Tis-
cini, 2021). Ou s udy ollows a small numbe o eme ging s udies
ha ocused on explo ing how a pa icula echnology o digi iza ion
suppo s he in e na ionaliza ion ac i i ies o fi ms (Hosseini, Fallon,
Wee akkody & Si a ajah, 2019;Shamim, Zeng, Choksy & Sha iq,
2020). Th ough a deep unde s anding o AI, ou s udy sheds ligh on
Table 4
Robus ness esul s.
Va iables P obi model Al e na i e measu e
Model 1 Model 2 Model 3 Model 4 Model 5
S a e owne ship 0.01 0.08*** 0.01* 0.01 0.02
(0.00) (0.03) (0.00) (0.00) (0.01)
Poli ical a fini y 0.74* 0.68* 4.53*** 0.76* 1.25*
(0.40) (0.40) (1.23) (0.40) (0.74)
Risk p e e ences o execu i es 0.32* 0.30* 0.29* 0.72** 0.46
(0.17) (0.17) (0.17) (0.34) (0.36)
O e seas expe ience 0.76*** 0.78*** 0.79*** 0.77*** 1.36***
(0.14) (0.14) (0.14) (0.14) (0.28)
Fi m size 0.23*** 0.24*** 0.25*** 0.22*** 0.36***
(0.06) (0.06) (0.06) (0.06) (0.11)
Fi m age 0.04*** 0.05*** 0.05*** 0.04*** 0.08***
(0.01) (0.02) (0.01) (0.01) (0.03)
Deb a io 1.36*** 1.34*** 1.34*** 1.37*** 1.43
(0.46) (0.46) (0.46) (0.46) (0.92)
Financial pe o mance 1.93 1.97 1.26 1.90 0.40
(1.27) (1.27) (1.26) (1.27) (2.32)
Tu no e 0.04 0.03 0.02 0.03 0.11
(0.14) (0.14) (0.14) (0.14) (0.26)
Ins i u ional dis ance 0.20*** 0.20*** 0.20*** 0.20*** 0.32***
(0.05) (0.05) (0.05) (0.05) (0.10)
Economic dis ance 0.30*** 0.30*** 0.32*** 0.30*** 0.50**
(0.11) (0.11) (0.11) (0.11) (0.22)
AI echnology 0.27*** 0.25*** 0.28*** 0.26***
(0.06) (0.06) (0.06) (0.06)
AI le el £S a e owne ship 0.01***
(0.00)
AI le el £Poli ical a fini y 0.54***
(0.16)
AI le el £Risk p e e ences o execu i es 0.15***
(0.05)
AI echnology (Al e na i e measu e) 0.63***
(0.23)
Yea dummies Yes Yes Yes Yes Yes
Cons an 1.20 1.73 0.99 0.85 1.72
(1.15) (1.19) (1.24) (1.14) (2.23)
Pseudo R
2
0.13 0.13 0.14 0.13 0.12
Obse a ions 654 654 654 654 539
No e: S anda d e o s a e in pa en heses; *** p<0.01, ** p<0.05, * p<0.1.
W. Liu, M. Cao, J. Zheng e al. Jou nal o Inno a ion & Knowledge 9 (2024) 100518
7
how he de elopmen o specific digi al echnologies can influence
fi ms’in e na ionaliza ion ac i i ies as well as hei s a egic deci-
sions (Meye e al., 2023).
Schola s ha e clea ly ecognized he ole o AI in business wi h
enhanced in o ma ion p ocessing and analy ical capabili ies ha can
ex end human cogni i e abili ies and suppo business decisions (Ja -
ahi, 2018). Specifically in he field o in e na ional business, AI can
help o ganiza ions achie e as e expansion o expo s and o e seas
ma ke s (Kopalle e al., 2022;Yang e al., 2023). In his s udy, we
ocus on he en y modes o MNEs in en e ing o eign ma ke s, seek-
ing o explain he an eceden mechanisms o his s a egic choice
based on he exis ing unde s anding o AI. We a gue ha echnologi-
cal capabili ies p o ide he basis o fi ms o achie e he e ogenei y
(Ede ing on & McCalman, 2008), while a he same ime unde mining
hos coun y unce ain y (Li & Xiong, 2022). Consis en wi h he p io
li e a u e on MNE s a egy, ou findings sugges ha fi ms a e mo e
willing o make highe esou ce commi men s in in e na ional
expansion when he echnological capabili ies hey possess c ea e
compe i i e ad an ages (B ou he s, 2002;Li & Xiong, 2022;Wei,
Zheng, Liu & Lu, 2014).
We also find e idence on he bounda y condi ions, aiming o be -
e unde s and he ela ionship be ween AI echnology and en y
mode o o eign ma ke s. Specifically, a highe sha e o s a e owne -
ship weakens he posi i e impac o AI echnology because s a e
owne ship o en implies con ol by he home go e nmen and is
associa ed wi h na ional secu i y. In addi ion, he highe he poli ical
a fini y be ween he wo coun ies, he mo e fi ms can exe a posi-
i e influence o AI echnology on he decision o en e wholly-
owned modes. This posi i e influence is also enhanced by he isk
p e e ences o execu i es, wi h isk-a e se execu i es being mo e
confiden and willing o make g ea e commi men s han isk-a e se
execu i es.
Theo e ical con ibu ions
Ou s udy con ibu es o he li e a u e on digi iza ion and in e na-
ional business. Digi aliza ion plays an impo an ole in in e na ional
business (Meye e al., 2023). Schola s ha e p e iously ocused on
he impac o digi iza ion as a whole on en y mode (Ekeledo & Si a-
kuma , 2004a;Li e al., 2021), bu ha e no specified how specific
echnologies o digi iza ion suppo fi ms’en y mode s a egies. We
build on he s udy o Reim e al. (2022)) in explo ing he impac o AI
as an impo an digi al echnology on en y mode. Ou s udy comple-
men s he di ec ion o s udies on AI echnologies o MNEs, whe e
p e ious discussions ha e ocused on he ole o AI in suppo ing
expo s, ma ke ing, human esou ces, and pe o mance (Denicolai e
al., 2021;Kopalle e al., 2022;Meye e al., 2023;Yang e al., 2023).
This s udy also con ibu es o he ole o AI echnology in knowl-
edge managemen . AI can help o ganiza ions o sha e aci knowl-
edge, con e aci knowledge in o explici knowledge, and analyze
Table 5
Resul s o he endogenei y es .
Va iables S age One S age Two
Model 1 Model 2 Model 3 Model 4 Model 5
Densi y o fibe op ic cables 0.02***
(0.01)
Densi y o mobile base s a ion 0.20***
(0.03)
S a e owne ship 0.02*** 0.01 0.02 0.01 0.01
(0.01) (0.01) (0.10) (0.01) (0.01)
Poli ical a fini y 1.13 0.75* 0.78* 4.39*** 0.98**
(0.70) (0.41) (0.42) (1.45) (0.45)
Risk p e e ences o execu i es 1.17*** 0.22 0.24 0.19 1.79*
(0.28) (0.22) (0.20) (0.22) (1.05)
O e seas expe ience 0.46* 0.79*** 0.79*** 0.82*** 0.96***
(0.24) (0.14) (0.14) (0.14) (0.20)
Fi m size 0.13 0.20*** 0.20*** 0.21*** 0.24***
(0.11) (0.06) (0.06) (0.06) (0.07)
Fi m age 0.08*** 0.04** 0.04** 0.04*** 0.06***
(0.02) (0.02) (0.02) (0.02) (0.02)
Deb a io 1.72** 1.35*** 1.39*** 1.31*** 1.56***
(0.85) (0.51) (0.52) (0.51) (0.56)
Financial pe o mance 0.82 2.14 2.22 1.48 2.57*
(2.51) (1.35) (1.36) (1.36) (1.49)
Tu no e 0.01 0.04 0.04 0.02 0.06
(0.26) (0.15) (0.15) (0.15) (0.18)
Ins i u ional dis ance 0.15* 0.18*** 0.18*** 0.19*** 0.22***
(0.09) (0.05) (0.05) (0.05) (0.06)
Economic dis ance 0.79*** 0.31** 0.30** 0.34*** 0.29**
(0.19) (0.13) (0.13) (0.13) (0.12)
AI echnology 0.33*** 0.36*** 0.34*** 0.47**
(0.09) (0.12) (0.09) (0.20)
AI le el £S a e owne ship 0.01
(0.01)
AI le el £Poli ical a fini y 0.52***
(0.20)
AI le el £Risk p e e ences o
execu i es
0.32**
(0.15)
Yea dummies Yes Yes Yes Yes Yes
Cons an 6.30*** 0.82 0.84 1.33 0.78
(2.17) (1.23) (1.38) (1.45) (1.21)
Obse a ions 615 615 615 615 615
No e: S anda d e o s a e in pa en heses; *** p<0.01, ** p<0.05, * p<0.1.
W. Liu, M. Cao, J. Zheng e al. Jou nal o Inno a ion & Knowledge 9 (2024) 100518
8
