Ra hnayake, A juna S ilal; T uong Dang Hoang Nha Nguyen; Ahn, Yonghan
A icle
Fac o s in luencing AI cha bo adop ion in go e nmen
adminis a ion: A case s udy o S i Lanka's digi al
go e nmen
Adminis a i e Sciences
P o ided in Coope a ion wi h:
MDPI – Mul idisciplina y Digi al Publishing Ins i u e, Basel
Sugges ed Ci a ion: Ra hnayake, A juna S ilal; T uong Dang Hoang Nha Nguyen; Ahn, Yonghan
(2025) : Fac o s in luencing AI cha bo adop ion in go e nmen adminis a ion: A case s udy o S i
Lanka's digi al go e nmen , Adminis a i e Sciences, ISSN 2076-3387, MDPI, Basel, Vol. 15, Iss. 5,
pp. 1-29,
h ps://doi.o g/10.3390/admsci15050157
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/321301
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/
Recei ed: 27 Feb ua y 2025
Re ised: 27 Ma ch 2025
Accep ed: 3 Ap il 2025
Published: 25 Ap il 2025
Ci a ion: Ra hnayake, A. S., Nguyen,
T. D. H. N., & Ahn, Y. (2025). Fac o s
In luencing AI Cha bo Adop ion in
Go e nmen Adminis a ion: A Case
S udy o S i Lanka’s Digi al
Go e nmen . Adminis a i e Sciences,
15(5), 157. h ps://doi.o g/10.3390/
admsci15050157
Copy igh : © 2025 by he au ho s.
Licensee MDPI, Basel, Swi ze land.
This a icle is an open access a icle
dis ibu ed unde he e ms and
condi ions o he C ea i e Commons
A ibu ion (CC BY) license
(h ps://c ea i ecommons.o g/
licenses/by/4.0/).
A icle
Fac o s In luencing AI Cha bo Adop ion in Go e nmen
Adminis a ion: A Case S udy o S i Lanka’s Digi al Go e nmen
A juna S ilal Ra hnayake 1,* , T uong Dang Hoang Nha Nguyen 2and Yonghan Ahn 3,*
1Depa men o Applied A i icial In elligence, Hanyang Uni e si y, E ica Campus,
Ansan-si 15588, Republic o Ko ea
2Cen e o AI Technology in Cons uc ion, Hanyang Uni e si y, E ica Campus,
Ansan-si 15588, Republic o Ko ea; nha [email p o ec ed]
3Depa men o A chi ec u e and A chi ec u al Enginee ing, Hanyang Uni e si y, E ica Campus,
Ansan-si 15588, Republic o Ko ea
*Co espondence: [email p o ec ed] (A.S.R.); [email p o ec ed] (Y.A.)
Abs ac : This s udy in es iga es he ac o s in accep ance o a i icial in elligence (AI)-
based cha bo applica ion in S i Lanka’s go e nmen adminis a ion se ices, which can
be applied o de eloping coun ies, using an ex ended echnology accep ance model (ex-
ended TAM) as a new esea ch amewo k by adding ex e nal cons uc s such as us ,
applica ion design/appea ance, and social in luence o he echnology accep ance model
(TAM). Conside ing he sus ainable implemen a ion o AI, i is c i ical o unde s and use
pe spec i es gi en he expanding and in ica e in eg a ion o AI echnology in go e nmen
ope a ions. Based on p e ious esea ch, his s udy p o ides a s uc u ed su ey o ind
ou esponden s’ hough s on using AI cha bo s o enhance go e nmen se ice deli e y.
Wi h a alid sample size o 207 esponses ob ained om S i Lanka, he da a we e analyzed
using a co a iance-based s uc u al equa ion model (CB-SEM) o es he hypo hesized
ela ionships. The indings e ealed ha social in luence (SI) has a posi i e and signi ican
impac on us (TR). Also, us and applica ion design (AD) ha e a posi i e and signi ican
impac on pe cei ed ease o use (PE), which in u n posi i ely in luenced pe cei ed use ul-
ness (PU) and hen PE posi i ely in luenced a i ude (AT) owa d beha io al in en ion (BI)
o accep AI cha bo applica ions in go e nmen adminis a i e se ices. The e o e, his
new model p o ed he e ec o new ex e nal ac o s and highligh s he impo ance o hose
ac o s in policy implemen a ions o u u e AI-d i en digi al go e nmen ini ia i es.
Keywo ds: AI cha bo s; digi al go e nmen ; echnology accep ance model; sus ainable AI
adop ion; use accep ance; us ; e-go e nmen
1. In oduc ion
A i icial in elligence echnologies, especially cha bo s, a e being used mo e and
mo e in he go e nmen and p i a e sec o s o enhance ope a ional e iciency and se ice
deli e y o suppo sus ainable go e nmen adminis a ion. Wi h he use o hese mode n
echnologies, agencies a e able o manage high olumes o ex e nal enqui ies, o e 24/7
suppo , and enhance use expe iences by ins an ly eplying o enqui ies (Chen e al.,
2023). To mee ci izens’ expec a ions o e ec i e go e nmen adminis a ion and u he
digi al go e nmen ini ia i es in S i Lanka, i is impe a i e ha AI be in eg a ed in o
go e nmen o ganiza ions. AI is ans o ming go e nmen s by op imizing decision making
p ocedu es, boos ing cus ome sa is ac ion, and cu ing down on adminis a i e wo k.
Go e nmen s use AI echnologies o examine eno mous olumes o da a, spo ends, and
Adm. Sci. 2025,15, 157 h ps://doi.o g/10.3390/admsci15050157
Adm. Sci. 2025,15, 157 2 o 29
decide on policies ha will be e se e he in e es s o he populace (Chiancone,2023). This
skill is essen ial o sol ing complica ed socie al issues since a i icial in elligence migh
p o ide insigh s ha con en ional analy ical echniques can be lacking. AI has a wide
ange o possible uses, om public sa e y and a ic managemen o heal hca e, whe e i
can speed up d ug disco e y and g ea ly imp o e ope a ional e iciency and esponse o
ci izen equi emen s.
On he o he hand, he use o AI in go e nmen unc ions aises signi ican e hical,
accoun abili y, and anspa ency issues. O ganiza ions in he go e nmen sec o mus
nego ia e he challenges o e hically implemen ing AI echnologies while gua an eeing ha
hese pla o ms p ese e p inciples like equi y and inclusi i y. Go e nmen s may au oma e
epe i i e ques ions, imp o e ci izen in e ac ions, and s eamline in e nal p ocedu es o
employees by u ilizing gene a i e AI. Bu o educe challenges and di ec he mo al use o
AI in e-go e nmen ini ia i es, success depends on c ea ing s ong amewo ks, educa ing
s a , and es ablishing go e nance s anda ds. Se ing esponsible p ac ices as a op p io i y
will be essen ial o building public us and op imizing he ad an ages o hese game-
changing echnologies as he ole o AI in he go e nmen de elopmen s.
When i comes o go e nmen sec o o ganiza ions adop ing AI cha bo s, pu ing
new echnologies in o p ac ice can be qui e di icul and expensi e (Hillemann,2023).
Many go e nmen s a e s uggling wi h o achie e he ad an ages ha AI echnology
and e-go e nmen ini ia i es a e expec ed o yield (Medaglia & Tangi,2022;Mikale
e al.,2023). Signi ican inancial losses ha e been caused by unsuccess ul echnological
implemen a ions, especially in he go e nmen sec o , unde sco ing he need o p edic ing
o ganiza ional and use needs. Low adop ion o e-go e nmen solu ions, like AI cha bo s,
is s ill a majo obs acle, limi ing bo h he physical and in angible bene i s, e en wi h he
po en ial o de elopmen (Chen e al.,2023). The success ul adop ion o hese echnologies
elies hea ily on use app o al and accep ance. Use inpu , which is equen ly es ic ed
in i s abili y o assess echnological iabili y, aids in he imp o emen o AI sys ems. The
possibili y o success ul adop ion can also be inc eased by ga he ing addi ional in o ma ion
and p ojec ions. I is possible o de e mine i a ce ain echnology, such as AI cha bo s,
will be success ully inco po a ed in o go e nmen sec o ope a ions by examining use
a i udes and beha io al in en s (Yigi canla e al.,2024).
O e he las decade, he e has been subs an ial de elopmen in unde s anding use
adop ion o new in o ma ion echnologies, pa icula ly wi h he help o he TAM (F. D.
Da is,1989). As a eliable amewo k o e alua ing use adop ion o de eloping echnolo-
gies, his model has ecei ed bo h heo e ical and expe imen al con i ma ion. Fo be e
explaining echnology adop ion, se e al schola s ha e expanded on TAM and p oposed
ex ended models ha include u he a iables. These imp o ed models gi e echnical
eams di ec ions o op imize sys em design and allow decision-make s o assess new ech-
nical se ices. Despi e i s Ame ican o igins, TAM has been shown in nume ous s udies
(Alalwan e al.,2018;Alenazy e al.,2019;Hsu & Lu,2004;Sai e al.,2024) o be a alid
model o desc ibing he ela ionship be ween use s and echnology accep ance in a a ie y
o scena ios.
Though TAM has been iden i ied as he mos sui able heo e ical ounda ion o iden-
i y he signi icance o use beha io and accep ance o mode n echnologies, he lack o
ex ended TAM applica ions in go e nmen AI adop ion can be explained as one o iden i-
ied esea ch gaps. Mos o he exis ing s udies on echnology adop ion in e-go e nmen
ely on TAM o Uni ied Theo y o Accep ance and Use o Technology (UTAUT) models
wi hou inco po a ing addi ional cons uc s (Venka esh & Bala,2008). Howe e , eme g-
ing esea ch sugges s ha ac o s like us , applica ion design/appea ance, and social
in luence a e c i ical in AI-based applica ions (Kelly e al.,2023;Om ani e al.,2022). So,
Adm. Sci. 2025,15, 157 3 o 29
his s udy ex ends TAM by inco po a ing hese ac o s, add essing his heo e ical gap.
Also, he e a e many esea ch s udies ha ha e been ca ied ou on cha bo adop ion in
p i a e sec o s such as e-comme ce, heal hca e, and banking. In addi ion, simila esea ch
s udies ocusing on go e nmen sec o s emain less common o none exis en (Zuide wijk
e al.,2021). Go e nmen sec o AI adop ion di e s due o ac o s such as adminis a i e
s uc u es, egula o y limi a ions, and ci izen con idence conce ns. The e o e, his esea ch
s udy is ying o ill his gap by ocusing speci ically on AI cha bo adop ion wi hin S i
Lanka’s digi al go e nmen se ices. Fu he mo e, mos o hose p e ious s udies on AI
cha bo adop ion a e ocused on de eloped coun ies and echnologically ad anced en i-
onmen s (Huang & Rus ,2018;Kasilingam & K ishna,2022). The e is a lack o empi ical
esea ch on AI adop ion in de eloping coun ies, pa icula ly in Sou h Asia. The e o e,
his s udy p o ides a S i Lankan pe spec i e, con ibu ing speci ic insigh s and add essing
he gap in non-de eloped, minimum echnologically ad anced coun ies in e-go e nmen
adop ion s udies.
The ollowing s udies ha e ex ended TAM o explo e echnology adop ion in e-
go e nmen . Howe e , hese s udies o en ocus on gene al e-go e nmen se ices, a he
han AI-based e-go e nmen cha bo s. An a icle on AI cha bo adop ion (Gopina h &
Kasilingam,2023) ocused on he comme cial sec o and i was no in a go e nmen ad-
minis a ion con ex . Ano he s udy (Sha ee e al.,2011) ocused on gene al e-go e nmen
se ices, no AI cha bo s in he con ex o e-go e nmen se ices. An (Sha ma & Aga wal,
2024) a icle on AI cha bo adop ion ocused on he educa ion sec o , and i did no conside
applica ion design. Addi ionally, an a icle on AI cha bo s in cus ome se ice (Kunz &
Wi z,2023) examined p i a e sec o cha bo s, no go e nmen sec o ones.
Ou s udy sugges s a unique TAM ex ension by in eg a ing us , applica ion de-
sign/appea ance, and social in luence in he con ex o AI cha bo s o public adminis a-
ion se ices in S i Lanka o analyze and unde s and how he use s’ pe cep ions o he
accep ance o AI cha bo applica ions in go e nmen o ganiza ions in luence hei sa is ac-
ion wi h he use o hese AI echnologies. To achie e his objec i e, a s uc u al equa ion
modeling app oach was applied, a s a is ical echnique ha allows he simul aneous e al-
ua ion o mul iple ela ionships be ween unobse able la en a iables. This app oach is
e ec i e o explo ing how he men ioned ac o s a ec use in en ions o use AI cha bo s
in go e nmen adminis a ion se ices. By ocusing on how he newly iden i ied ex e nal
cons uc s in luence use beha io and ocusing on ac ual AI cha bo accep ance, his s udy
aims o o e a new model ha can guide go e nmen s and go e nmen policy make s
o design a sus ainable echnology accep ance plan o success ul e-go e nmen se ice
deli e y o de eloping coun ies.
Cu en A ailabili y o E-Go e nmen Se ices in S i Lanka
S i Lanka has succeeded in making conside able p og ess in he implemen a ion o
e-go e nmen se ices o e he pas en yea s. The S i Lankan go e nmen has ecognized
he po en ial o digi al ans o ma ion o enhance go e nmen e iciency, e ec i eness, and
imp o e ci izens’ access o day- o-day se ices. Howe e , he in ensi y o digi al en i on-
men adop ion and he a ailabili y o e-se ices a y ac oss di e en go e nmen sec o
o ganiza ions and se ices. The S i Lankan go e nmen has ini ia ed some e-go e nmen
pla o ms o p o ide e-se ices o ci izens. One o he key ini ia i es is he S i Lanka Go -
e nmen Po al, which se es as a cen al hub o accessing a ious e-se ices. Key se ices
such as ax iling, ehicle e enue licensing, online appoin men ese a ion, and paymen
o u ili y bills a e a ailable h ough hese digi al pla o ms. Howe e , he use o AI cha bo s
in public adminis a ion is s ill in i s nascen s ages, and no go e nmen agencies ha e ully
implemen ed hese e-sys ems o suppo adminis a i e se ices o ci izens.
Adm. Sci. 2025,15, 157 4 o 29
When we conside he ac ual usage and ci izen adop ion o hose e-se ices, i a ied
based on he geog aphical and de elopmen a ea. U ban and de eloped a eas, whe e access
o he in e ne se ice and digi al in as uc u e is be e , ha e seen sophis ica ed usage
compa ed o u al a eas, whe e he e a e issues such as limi ed in e ne se ice connec i i y
and digi al li e acy. T adi ional me hods such as in-pe son isi s o go e nmen o ices and
pape -based documen p ocessing a e con inuing o lead in many go e nmen o ices in
unde eloped u al a eas. Al hough he S i Lankan go e nmen pu e o s o digi alize
go e nmen se ices, some ci izens and public sec o employees a e ei he un amilia ,
esis an o change, o lack o us in e-go e nmen pla o ms, esul ing in lowe usage
and adop ion o e-se ices.
2. Theo e ical F amewo k and Hypo heses De elopmen
The accep ance o AI echnological applica ion is complica ed due o i s s uc u e,
and he TAM alone canno be a comp ehensi e ool o his. TAM needs o be me ged
wi h o he signi ican cons uc s ha guide he plan o a new model, compa ible wi h
la es AI echnological aspec s such as cha bo applica ions. In his s udy, e ol ing and
accumula ing a comp ehensi e lis o cus ome beha io de e minan s o sus ainable AI
echnology accep ance is o assess he impac o hese indica o s. The cus ome can decide
whe he o ake he igh in e en ions o no o maximize he e ec i e u iliza ion o he new
ansac ion echnology. Howe e , he AI cha bo applica ion is new and has complica ed
cha ac e is ics in e ms o adop ion and de elopmen in go e nmen sec o o ganiza ions.
Th ee iden i ied cons uc s may play di ec and indi ec oles in he sus ainable adop ion o
AI cha bo applica ion, ema kably in he go e nmen adminis a ion se ice ini ia i es in
he S i Lankan con ex ( us , applica ion design/appea ance, and social in luence).
2.1. Technology Accep ance Model (TAM)
The TAM (Figu e 1) has been used in esea ch o explo e he accep ance o new
echnology o new se ices (F. D. Da is,1989;F. D. Da is & Venka esh,1996). TAM is one
o he mos e ec i e con ibu ions o Ajzen and Fishbein’s heo y o easoned ac ion (TRA).
Da is’s echnology accep ance model, TAM (F. D. Da is,1989;F. D. Da is & Venka esh,
1996), is he mos widely u ilized model o accep ance and usage o inno a i e echnology
by use s.
Adm. Sci. 2025, 15, x FOR PEER REVIEW 4 o 30
and paymen o u ili y bills a e a ailable h ough hese digi al pla o ms. Howe e , he
use o AI cha bo s in public adminis a ion is s ill in i s nascen s ages, and no go e nmen
agencies ha e ully implemen ed hese e-sys ems o suppo adminis a i e se ices o
ci izens.
When we conside he ac ual usage and ci izen adop ion o hose e-se ices, i a ied
based on he geog aphical and de elopmen a ea. U ban and de eloped a eas, whe e ac-
cess o he in e ne se ice and digi al in as uc u e is be e , ha e seen sophis ica ed us-
age compa ed o u al a eas, whe e he e a e issues such as limi ed in e ne se ice con-
nec i i y and digi al li e acy. T adi ional me hods such as in-pe son isi s o go e nmen
offices and pape -based documen p ocessing a e con inuing o lead in many go e nmen
offices in unde eloped u al a eas. Al hough he S i Lankan go e nmen pu effo s o
digi alize go e nmen se ices, some ci izens and public sec o employees a e ei he un-
amilia , esis an o change, o lack o us in e-go e nmen pla o ms, esul ing in lowe
usage and adop ion o e-se ices.
2. Theo e ical F amewo k and Hypo heses De elopmen
The accep ance o AI echnological applica ion is complica ed due o i s s uc u e,
and he TAM alone canno be a comp ehensi e ool o his. TAM needs o be me ged
wi h o he signi ican cons uc s ha guide he plan o a new model, compa ible wi h la -
es AI echnological aspec s such as cha bo applica ions. In his s udy, e ol ing and ac-
cumula ing a comp ehensi e lis o cus ome beha io de e minan s o sus ainable AI
echnology accep ance is o assess he impac o hese indica o s. The cus ome can decide
whe he o ake he igh in e en ions o no o maximize he effec i e u iliza ion o he
new ansac ion echnology. Howe e , he AI cha bo applica ion is new and has compli-
ca ed cha ac e is ics in e ms o adop ion and de elopmen in go e nmen sec o o gani-
za ions. Th ee iden i ied cons uc s may play di ec and indi ec oles in he sus ainable
adop ion o AI cha bo applica ion, ema kably in he go e nmen adminis a ion se ice
ini ia i es in he S i Lankan con ex ( us , applica ion design/appea ance, and social in-
luence).
2.1. Technology Accep ance Model (TAM)
The TAM (Figu e 1) has been used in esea ch o explo e he accep ance o new ech-
nology o new se ices (Da is, 1989; Da is & Venka esh, 1996). TAM is one o he mos
effec i e con ibu ions o Ajzen and Fishbein’s heo y o easoned ac ion (TRA). Da is’s
echnology accep ance model, TAM (Da is, 1989; Da is & Venka esh, 1996), is he mos
widely u ilized model o accep ance and usage o inno a i e echnology by use s.
Figu e 1. TAM.
Figu e 1. TAM.
Use s’ pe cep ions abou he ac ual u ili y o echnology (ac ual sys em use) we e
ound o be ela ed o hei a i ude and beha io al in en ion o use he echnology. Pe -
cei ed use ulness exhibi s a mo e ha monious associa ion wi h u iliza ion han he o he
model a iables. As a esul , his s udy decides o inco po a e pe cei ed use ulness and
pe cei ed ease o use in o a new s udy pa adigm. Pe cei ed use ulness is de ined as he
ex en o which a use belie es ha adop ing a ce ain sys em will imp o e wo k pe o -
Adm. Sci. 2025,15, 157 5 o 29
mance. Pe cei ed ease o use is he deg ee o which a use belie es ha u ilizing a speci ic
sys em will be e o less and easily adop ed.
2.2. Hypo hesis and Model De elopmen
Acco ding o he p o en li e a u e, TAM has yielded good esul s o calcula ing he
beha io al in en ion o use he new echnology (Edo e al.,2023;Ikhsan e al.,2025;Na asia
e al.,2022;Sai e al.,2024). Howe e , he e is a lack o a TAM o ob ain good esul s
ha alue he accep ance o he la es ad anced echnologies, and he de elopmen o his
ex ended TAM is equi ed o achie e his app oach. Fu he mo e, we iden i ied h ee new
ac o s ha a e no explained in he TAM ( us , applica ion design/appea ance, social
in luence). Also, hese ac o s a e conside ed impo an acco ding o many expe s in AI
echnology and he cha ac e is ics and unique s uc u e o his echnology.
This pape will in oduce a new esea ch model wi h wo pa s and will explain hem
in de ail wi h each hypo hesis: i s , he undamen al TAM cons uc s (beha io in en ion,
a i ude, pe cei ed use ulness, pe cei ed ease o use) and second, ex e nal cons uc s ( us ,
applica ion design/appea ance, and social in luence).
2.2.1. TAM-Fundamen al Cons uc s
Beha io In en ion (BI)
Acco ding o beha io al psychology, use beha io is in luenced by in en ionali y,
i also ela es o he use ’s pe cei ed possibili y o p obabili y ha hey will engage in a
speci ic beha io , in his case, expe iencing he new echnology (Malho a & Galle a,1999).
Beha io al in en ion aids in he iden i ica ion o well- o med measu es o use accep ance
ea ly in he sys em de elopmen li e cycle. Fu he mo e, i assis s clien s in accep ing
help ul inno a ions o ejec ing w ong and ha m ul ones, hence educing he dange o
gi ing in e io echnologies p io o ejec ion (F. Da is e al.,1989;F. D. Da is,1989).
Ano he s udy sugges ed ha he Pe cei ed use ulness o a echnology di ec ly in-
luences use s’ beha io al in en ion o use i (Ilyas e al.,2023). Speci ically, when use s
pe cei e echnology as mo e bene icial, hey a e mo e likely o accep and u ilize i . Be-
ha io al in en ion e e s o a use ’s subjec i e mo i a ions o pe o ming a beha io on
a sys em. The in en ion o use such a sys em is d i en by use pu pose beha io (Wa -
shaw & Da is,1985). Also, ano he au ho ound ha pe cei ed mo al alues ac as a
unique pe o mance o beha io al in en ion (Ajzen & Fishbein,1970). Pe haps he use
in luences beha io al in en ions based on hei a i udes and alues in such a mo ali y
scena io (Go such & O be g,1983). Ano he a icle indica ed ha habi was a mo e po-
en p edic o o class oom beha io han in en ions. Howe e , a pos - esea ch analysis
suppo ed he idea ha in en ions become impo an when he habi componen can be
supp essed (Landis e al.,1978). The a i ude and in en ion ela ionship was a enua ed
when he ex en o pas beha io was included as an explana o y a iable. Simila ly, pas
beha io lessened he impac o in en ions on beha io (Bagozzi,1981). The e is e idence
linking se e al dimensions o beha io , and he e is a co ela ion be ween cogni i e and
beha io al measu emen s, which we e de ined mo e by chance han by o mal logic (Wa -
shaw,1980). Ano he au ho e ealed ha addi ional s udy is equi ed o examine PU and
PE om a b oad pe spec i e in o de o calcula e he impac o ou side a iables on hese
in e nal beha io al indings (
F. D. Da is,1989
). In en ionali y in beha io is he ul ima e
objec i e. As men ioned abo e, nume ous s udies ha e iden i ied measu es o a i ude
and hen calcula ed how closely hey co ela e wi h beha io . A mo e logical cou se o
ac ion would be o concen a e on beha io measu emen and p edic ion, which would
lead o he iden i ica ion o he p ima y ac o s and cha ac e is ics in luencing consume s’
beha io al in en ion o emb ace AI cha bo echnology. The e o e, he use beha io al
Adm. Sci. 2025,15, 157 6 o 29
in en ion o accep AI cha bo applica ion in he S i Lankan e-go e nmen mo emen needs
o be iden i ied and weigh ed when analyzing such enhanced model.
A i ude (AT)
Rega ding in en ionali y, he e m “a i ude” desc ibes how a use eels abou he
new echnology, whe he posi i ely o nega i ely (F. Da is e al.,1989;F. D. Da is,1989).
The heo y o easoned ac ion (TRA) (Fishbein & Ajzen,1975) led esea che s o ind
ha he ac ual beha io , an a i ude owa ds u ilizing and in es iga ing objec s such a
echnological sys em, is e e ed o as he use belie sys em. When de e mining hei
beha io al in en ions, people conside hei a i udes owa ds each o he a ailable op ions.
I seems ha he a i ude owa ds simila choice p ocesses does no e eal how an indi idual
o ms hei opinions abou whe he o no o ca y ou se e al asks (Sheppa d e al.,1988).
Conside ing desi es and he causes o he u ge could a ec he salien esul o he goal
beha io . The no ma i e belie s and a i ude alues can be me ged unde he a e age
o an icipa ion (F. Da is,1985;F. Da is e al.,1989). Acco ding o se e al s udies, one’s
a i ude can in luence and be in luenced by he a i udes o o he s (Ajzen & Fishbein,2000,
2005). I has been assumed ha social in luence p ocessing is c ucial o a new sys em’s
accep abili y (Ajzen & Fishbein,2005). Fu he mo e, he use o new echnology i sel may
cause a i udes owa ds i o shi , which could di ec ly a ec o ganiza ional s uc u e,
communica ion s yles, and wo king loca ions (Cuel & Fe a io,2009;Rice & Aydin,1991).
Also, ano he s udy shows ha he a i ude and beha io al in en ion owa d a speci ic
applica ion is de e mined by he use ulness pe cep ion, while i s use ulness is in luenced
by se e al ex e nal ac o s (P abowo & Nug oho,2019;Wang e al.,2023). Simila ly, a
s ong co ela ion be ween a i ude and use beha io links is gua an eed when app op ia e
measu emen execu ion is used (Ajzen & Fishbein,1977). As a esul , he TAM and u he
s udy indings ha e alida ed he associa ion and consequences be ween a i ude and
beha io al in en ion. Based on hose alida ed concep s, i can be s a ed ha he a i ude
o ci izens owa d he accep ance o AI cha bo s will in luence hei beha io al in en ion
o use hese e-go e nmen sys ems in he u u e. A posi i e a i ude is necessa y o
adop ion, as i e lec s a willingness o engage wi h he new echnology. In S i Lanka, he
adop ion o digi al ini ia i es in go e nmen se ices has been slow due o conce ns abou
e iciency and us wo hiness. Howe e , as AI-powe ed cha bo s p omise o s eamline
go e nmen se ices, ci izens’ posi i e a i udes owa ds he pe cei ed bene i s can inc ease
hei in en ion o use hese se ices. Ci izens who ind he cha bo use ul and easy o in e ac
wi h a e mo e likely o adop i . The e o e, he ollowing hypo hesis is made:
Hypo hesis 1 (H1): A i ude has a posi i e and signi ican impac on beha io al in en ion owa d
AI cha bo applica ion adop ion in go e nmen adminis a ion se ices.
Pe cei ed Use ulness (PU)
The ex en o which a use eels ha u ilizing he new echnology will imp o e his
o he pe o mance is known as pe cei ed use ulness (F. Da is,1985;F. D. Da is,1989).
This au ho disco e ed ha he a ge ed pa ial in e en ion could a ec a i udes and
belie s
(F. D. Da is,1989)
. Pe cei ed use ulness is s ongly in luenced by in en ion, bu
a i ude has a limi ed co ela ion wi h pe cei ed use ulness (Ma ia & Sugiyan o,2023). This
was cla i ied in his w i ings which su eyed he opic o people who wish o use help ul
echnology despi e ha ing a nega i e a i ude owa ds i . Theo e ically, PU is a posi i e
aspec ; use s a e mo e likely o suppo an applica ion based on i s pe o mance capabili ies
and abili ies han on how easy o ha d he sys em is o use, which a ec s se ice adop ion
(F. Da is e al.,1989). This sugges s ha he cha ac e is ics o pe cei ed use ulness a e
connec ed and a ec he deg ee o use (Adams e al.,1992).
Adm. Sci. 2025,15, 157 7 o 29
Use PU can be caused by a a ie y o a iables, such as en i onmen al ac o s,
ha ha e he po en ial o signi ican ly al e consume pe cep ions (Banja naho ,2017).
I is sugges ed ha he con ex ual ac o s o pe cei ed en i onmen al unce ain y and
decen aliza ion will a ec he PU o agg ega ed da a. The bene i s o he p io in eg a ion
o PU in hings like echnology sys ems a e explained by ce ain disco e ies (Saade &
Bahli,2005). F om ano he pe spec i e, ano he s udy ound ha PU and enjoymen had a
compa able e ec on he equency and ime o use; he e ec o compu e anxie y was mo e
abou enjoymen han pe cei ed use ulness (Igba ia e al.,1994). Since many componen s
o he sys em en i onmen and PU had a majo impac on PE o he echnology sys em,
he e was a di ec impac on he in o ma ion echnology sys em, pa icula ly on pe cei ed
ease o use (Ka ahanna & S aub,1999).
Conside ing hose p e ious s udies, he sugges ion is ha he use ulness o AI cha -
bo s will shape ci izens’ a i udes owa d hei use. S i Lankan ci izens o en expe ience
ine iciency in public se ice deli e y. I he cha bo applica ion is seen as imp o ing se ice
deli e y and add essing go e nmen se ice challenges, use s a e mo e likely o gain a
posi i e a i ude owa ds accep ing i . A use ul cha bo ha educes wai ing imes, p o ides
ins an eedback, and manages ou ine asks e ec i ely will gene a e a posi i e a i ude
owa d i s adop ion. Then, he ci izens will mo i a e hemsel o use he cha bo i hey
pe cei e i as an e ec i e solu ion o cu en se ice sho comings. The e o e, he ollowing
hypo hesis is made:
Hypo hesis 2 (H2): Pe cei ed use ulness has a posi i e and signi ican impac on a i udes owa ds
AI cha bo applica ion adop ion in go e nmen adminis a ion se ices.
Pe cei ed Ease o Use (PE)
The deg ee o which people belie e ha u ilizing he new echnology would be e o -
less is known as pe cei ed ease o use (F. Da is,1985;
F. Da is e al.,1989
). PE in he TAM
is one o he p ima y cons uc s. This cons uc has wo di ec cons uc i e e ec s on PU
and AT (F. Da is e al.,1989). Nume ous s udies ha e endo sed and employed TAM heo y
o es ima e consume beha io wi h new echnologies (F. Da is,1985;
F. Da is e al.,1989
).
PE is he likelihood ha use s will expec he in ended sys em o be e o less (G ani´c &
Ma anguni´c,2019;Ma hieson,1991). In cons uc ion, PE is he deg ee o which he use an-
icipa es and hinks ha u ilizing his se ice o echnical sys em will be e o less (F. Da is
e al.,1989;Nakisa e al.,2023). When a emp ing o con ey he emo ions and aspi a ions
o he clien s o he se ice p o ide s, c ea ing a sys em ha is simple, esponsi e, easy o
use, easy o manage, and adap able is c ucial o achie ing ou eally challenging objec i es
(Gould & Lewis,1985).
PE was aken in o conside a ion by many esea che s o de e mine use accep ance.
Addi ionally, e en hough all bu a ew o hese esea che s ecei ed he an icipa ed esul s
ega ding PE, he TAM was widely used in p ac ice in his ield (Venka esh,2000). The
signi icance o i s hand expe ience in loca ing he PE was alida ed by he indings o
ea lie use accep ance s udies ha ocused on ac ual use (Calisi & Calisi ,2004;Ge en &
S aub,2000;Hackba h e al.,2003;Saade & Bahli,2005;Venka esh & Da is,1996).
The PE o AI cha bo s will in luence ci izens’ a i udes owa d hem. I use s ind
he sys em easy o use, hei a i ude owa ds using i will be mo e posi i e. In S i Lanka,
digi al li e a eness is an impo an ac o in he accep ance o mode n echnology. Mos
subu ban S i Lankans may no ha e sound echnical skills, so a cha bo ha is simple and
use - iendly will posi i ely build hei a i udes owa d using i . As AI echnology is s ill
new o many a eas, ensu ing ease o use can imp o e echnological conce ns, leading o a
mo e a o able pe cep ion o he sys em. Also, i he cha bo is easie o use, i will be mo e
use ul. Then, ci izens will ind ha he cha bo is mo e e icien and easie and will be mo e
Adm. Sci. 2025,15, 157 8 o 29
likely o belie e in i and use i in go e nmen se ices. Ease o use is c i ical aspec in a
de eloping coun y like S i Lanka. Because a conside able pe cen age o he popula ion
has limi ed access o mode n echnological ools. I go e nmen cha bo applica ions a e
simple o expe ience, use s will pe cei e i as a use ul ool o engaging wi h S i Lankan
e-go e nmen se ices. The e o e, he ollowing hypo heses a e made:
Hypo hesis 3 (H3): Pe cei ed ease o use has a posi i e and signi ican impac on a i ude owa d
AI cha bo applica ion adop ion in go e nmen adminis a ion se ices.
Hypo hesis 4 (H4): Pe cei ed ease o use has a posi i e and signi ican impac on pe cei ed
use ulness owa d AI cha bo applica ion adop ion in go e nmen adminis a ion se ices.
2.2.2. Ex e nal Cons uc s
The ollowing ex e nal cons uc s a e adop ed acco ding o en i onmen al and ech-
nology cha ac e is ics. Mo eo e , some s udies ha ha e ensu ed hese de ini ions and
ela ionships a e shown in Table 1.
Table 1. De ini ions o ex e nal cons uc s.
Ex e nal Cons uc Gene al Concep ualiza ion Sou ce(s)
T us
T us in ol es he con idence ha use s place
in a echnology, belie ing ha i can pe o m
i s unc ions eliably and will ac in he use s’
bes in e es .
(Bu ke e al.,2007;Dhaga a e al.,2020;
Hasija & Espe ,2022;Hong,2025;
Ja enpaa e al.,2000)
Applica ion Design/
Appea ance
Applica ion design and appea ance e e o
he isual and unc ional elemen s o
echnology ha in luence use pe cep ions,
a ec ing usabili y, sa is ac ion,
and accep ance.
(Hoehle & Venka esh,2015;
Zhou e al.,2009)
Social In luence
Social in luence pe ains o he impac ha
pee s, amily, and la ge social ne wo ks ha e
on indi iduals’ a i udes and beha io s
owa d adop ing new echnologies.
(Chaouali e al.,2016;Cheng e al.,2022)
T us (TR)
T us is he deg ee o which consume s eel secu e, a ease, and con iden when
u ilizing echnology (Ja enpaa e al.,2000;McCloskey,2006). Fac o s ha ei he di ec ly
o indi ec ly mo i a e indi iduals o adop echnology include us , secu i y, and p i acy
(Ma emba & Li,2017). A us wo hy sys em ha adjus s o he una oidable changes in
us can manage he way social in e ac ions change o e ime (Golbeck & Ku e ,2009).
Indi iduals wi h posi i e opinions abou a gi en echnology may be mo e open o us and
eel mo e secu e abou i han hose wi h un a o able sen imen s. When us is aken in o
accoun , his demons a es s ong pa ial co ela ions be ween isk and accep ance (Eise
e al.,2006). Also, ano he esea ch s udy indica ed ha e-comme ce cha bo s enhance
us and accessibili y, encou aging use s o become less willing o ake chances and
be e p o ec ing hem agains he likelihood o un us wo hiness si ua ions (Celik e al.,
2022). In he case o new echnology-based applica ions, us should be high, and he isk
p obabili y should be educed. Bo h mindse and he PE o he a ailable echnology a e
di ec ly impac ed by us . T us is he key elemen o his s udy, and his ac o shows
a signi ican indi ec e ec on cus ome beha io . T us has he powe o in luence a
cus ome ’s choice o echnology o e en se ice (Kesha wani & Bish ,2012).
Adm. Sci. 2025,15, 157 15 o 29
nan alidi y (Fo nell–La cke c i e ion). Meanwhile, he s uc u al model was e alua ed
using pa h coe icien s and p- alues.
4. Resul s
The heo e ical model o his esea ch was based on he use o a co a iance-based
s uc u al equa ion modeling app oach. The da a analyzed in IBM SPSS AMOS (V.23)
esul ed in wo models: a measu emen model (see Figu e 3), which e alua es he eliabili y
o he cons uc s, alidi y, and model i and a s uc u al model, which add esses he
hypo heses, di ec -indi ec - o al e ec s ela ed o he success ul AI cha bo adop ion in
go e nmen adminis a ion se ices in S i Lanka.
Adm. Sci. 2025, 15, x FOR PEER REVIEW 15 o 30
cons uc s we e assessed h ough a measu emen model u ilizing a Con i ma o y Fac o
Analysis (CFA) wi h IBM SPSS AMOS (V23).
The me hodological echnique in ol ed e alua ing bo h measu emen and s uc u al
models. The s uc u al model examined la en a iable associa ions, whe eas he meas-
u emen model assessed eliabili y and alidi y. The e lec i e measu emen model was
e alua ed based on indica o loadings, in e nal consis ency eliabili y (C onbach’s alpha
and composi e eliabili y), con e gen alidi y (a e age a iance ex ac ed), and disc imi-
nan alidi y (Fo nell–La cke c i e ion). Meanwhile, he s uc u al model was e alua ed
using pa h coefficien s and p- alues.
4. Resul s
The heo e ical model o his esea ch was based on he use o a co a iance-based
s uc u al equa ion modeling app oach. The da a analyzed in IBM SPSS AMOS (V.23) e-
sul ed in wo models: a measu emen model (see Figu e 3), which e alua es he eliabili y
o he cons uc s, alidi y, and model i and a s uc u al model, which add esses he hy-
po heses, di ec -indi ec - o al effec s ela ed o he success ul AI cha bo adop ion in go -
e nmen adminis a ion se ices in S i Lanka.
Figu e 3. Measu emen model.
4.1. Measu emen Model Assessmen —Reliabili y, Validi y, and C oss Loadings
E e y measu emen i em needs o ha e i s alidi y, eliabili y, and ac o loading as-
sessed. A measu e’s consis ency is i s eliabili y. When a measu e yields consis en esul s
unde consis en ci cums ances, i is deemed us wo hy (J. Hai e al., 2022). Fo each
i em loading o be deemed eliable, he alue mus be equal o o g ea e han (0.5).
C onbach’s alpha (α) and composi e eliabili y (CR) a ings we e used o assess he con-
s uc ’s in e nal consis ency. C onbach’s alpha, wi h ecommended alues o 0.7 o 0.8,
measu es how well a se o objec s ep esen s a unidimensional la en concep . C onbach’s
alpha alues o all cons uc s exceeded he commonly accep ed h eshold o 0.70,
Figu e 3. Measu emen model.
4.1. Measu emen Model Assessmen —Reliabili y, Validi y, and C oss Loadings
E e y measu emen i em needs o ha e i s alidi y, eliabili y, and ac o loading
assessed. A measu e’s consis ency is i s eliabili y. When a measu e yields consis en esul s
unde consis en ci cums ances, i is deemed us wo hy (J. Hai e al.,2022). Fo each i em
loading o be deemed eliable, he alue mus be equal o o g ea e han (0.5). C onbach’s
alpha (
α
) and composi e eliabili y (CR) a ings we e used o assess he cons uc ’s in e nal
consis ency. C onbach’s alpha, wi h ecommended alues o 0.7 o 0.8, measu es how well
a se o objec s ep esen s a unidimensional la en concep . C onbach’s alpha alues o
all cons uc s exceeded he commonly accep ed h eshold o 0.70, indica ing sa is ac o y
in e nal consis ency. Composi e eliabili y, wi h ecommended alues g ea e han 0.7,
assesses he dependabili y o indica o s connec ed wi h a speci ic elemen . While bo h
me ics e lec in e nal consis ency, CR is p e e able o C onbach’s alpha o cons uc -
le el assessmen s in s uc u al equa ion modeling analysis. In his s udy, C onbach’s
alpha alues anged om 0.765 o 0.923, and CR alues anged om 0.850 o 0.946 o all
cons uc s, con i ming ha hey exhibi accep able in e nal consis ency.
Adm. Sci. 2025,15, 157 16 o 29
Then, cons uc alidi y was assessed using a con e gen alidi y echnique; he
a e age a iance ex ac ed (AVE), which is he g and mean alue o he squa ed loadings
o he i ems ele an o he cons uc , is he ypical me ic o p o ing con e gen alidi y.
Validi y is he deg ee o which a cons uc ’s indica o s join ly measu e. To his ex en , a
la en cons uc explains he a ia ion in i s indica o s. When he AVE alue is 0.5 o highe ,
i indica es ha he cons uc explains o e hal o he a ia ion in i s elemen s (J. Hai e al.,
2016). As desc ibed in Table 3, all AVE alues a e g ea e han 0.5. This p o es ha he
con e gen alidi y o he cons uc s o his s udy is sa is ied.
Table 3. Measu emen model ac o loadings, eliabili y, and in e nal consis ency.
Fac o Code Desc ip ion Loading AVE C.R C.A
T us
TR1 Cha bo applica ion is us wo hy 0.767
0.589 0.850
TR2
Cha bo applica ion p o ide s gi e he imp ession
ha hey keep p omises and commi men s on
in o ma ion p o ided
0.752 0.765
TR3 Cha bo applica ion p o ide s keep my bes
in e es s in mind. 0.854
TR4 Cha bo can add ess my issues 0.686
Applica ion
Design/
Appea ance
AD1
I will accep his cha bo s applica ion i he design
o be simila o o he sys ems ha I used o
know o .
0.698
0.716 0.909 0.864
AD2
I will accep his cha bo applica ion i he cha bo s
se ice applica ion is simple o na iga e. 0.898
AD3 I will accep his cha bo applica ion i i clea ly
gene a es and shows my equi ed esponse. 0.908
AD4 I will accep his cha bo applica ion i i ope a es
e ec i ely and ee om echnical issues. 0.865
Social
In luence
SI1 I will use his cha bo applica ion i he se ice is
widely used by people in my communi y. 0.658
0.704 0.903 0.858
SI2 I hink ha I will adop his cha bo applica ion i
my supe iso s/senio s use i . 0.862
SI3 I hink ha I will adop his cha bo applica ion i
my iends use i . 0.921
SI4 I will adop his cha bo applica ion i my amily
membe s/ ela i es use i . 0.889
Pe cei ed
Ease o Use
PE1 I hink lea ning o ope a e he cha bo applica ion
would be easy o me 0.789
0.683 0.866 0.767
PE2 I belie e i would be easy o ge he cha bo
applica ion o accomplish wha I wan o do. 0.841
PE3 I is easy o me o become skill ul a using his
cha bo applica ion. 0.848
Pe cei ed
Use ulness
PU1 Using his cha bo applica ion would imp o e he
quali y o public se ice. 0.861
0.771 0.931 0.900
PU2 Using his cha bo applica ion would inc ease
my p oduc i i y. 0.866
PU3
Using his cha bo applica ion would sa e ime on
ge ing go e nmen in o ma ion and se ices. 0.905
PU4 I belie e his cha bo applica ion is use ul o
deli e y o public se ices online o ci izens. 0.879
Adm. Sci. 2025,15, 157 17 o 29
Table 3. Con .
Fac o Code Desc ip ion Loading AVE C.R C.A
A i ude
AT1 I is a good idea o use a cha bo applica ion in he
public sec o . 0.889
0.815 0.946 0.923
AT2 I is wise o use a cha bo applica ion in he
public sec o . 0.907
AT3 I like o use a cha bo applica ion in he
public sec o . 0.930
AT4 I is pleasan o use a cha bo applica ion in
public sec o . 0.883
Beha io al
In en ion
BI1
I I ha e access o his cha bo applica ion, I in end
o use i . 0.878
0.740 0.895 0.817
BI2 I I ha e access o his cha bo applica ion, I will
use i . 0.899
BI3 I plan o use his cha bo applica ion wi hin he
nex 6 mon hs. 0.800
No e: AVE—a e age a iance ex ac ed; C.R—composi e eliabili y; C.A—C onbach’s alpha.
Fo u he analysis o cons uc alidi y, he disc iminan alidi y echnique will be
used in his s udy. Disc iminan alidi y uses he uni o mi y and alidi y o concen a ion
o con i m whe he each componen is impo an on i s own wi hou in e e ing wi h
o he ac o s. The Fo nell and La cke app oach is used o e alua e disc iminan alidi y,
which is a c ucial aspec o measu emen model eliabili y and alidi y (Fo nell & La cke ,
1981). This me hod compa es he squa e oo o he AVE o a cons uc o he co ela ions
be ween o he cons uc s. The diagonal alue mus be bigge han he co ela ions be ween
o he cons uc s. Acco ding o Table 4, he AVE o each cons uc exceeds he co ela ions
be ween ha cons uc and any o he cons uc in his s udy model. Fo example, TR exhib-
i ed a squa e oo o AVE o 0.767, which is signi ican ly highe han i s co ela ions wi h
o he cons uc s. I indica es ha he cons uc s explain mo e a iance in hei espec i e
i ems han hey sha e wi h o he cons uc s. The e o e, e en beyond he s anda d AVE
alues, he Fo nell and La cke c i e ion con i ms ha his measu emen model has s ong
disc iminan alidi y.
Table 4. Disc iminan alidi y is based on Fo nell and Lacke c i e ion.
TR AD SI PE PU AT BI
TR 0.767
AD 0.223 0.846
SI 0.356 0.341 0.839
PE 0.455 0.516 0.296 0.826
PU 0.323 0.440 0.227 0.652 0.878
AT 0.215 0.221 0.166 0.395 0.600 0.903
BI 0.258 0.553 0.217 0.544 0.658 0.401 0.860
No e. TR— us ; AD—applica ion design/appea ance; SI—social in luence; PE—pe cei ed ease o use;
PU—pe cei ed use ulness; AT—a i ude; BI—beha io al in en ion.
Also, Va iance In la ion Fac o (VIF) alues we e compu ed u ilizing Squa ed Mul iple
Co ela ions (SMC) acqui ed om AMOS o e alua e mul icollinea i y among he la en
componen s. The indings show ha all cons uc s’ VIF alues (TR = 1.25, PE = 1.57,
PU = 1.65
, AT = 1.63, BI = 1.22) a e signi ican ly below he sugges ed cu o o 5. VIF alues
less han 5 indica e ha mul icollinea i y is no an issue (J. F. Hai e al.,2019). As a esul ,
he cons uc s in his s udy a e independen , gua an eeing he objec i i y o he s uc u al
model’s p edic ion co ela ions.
Adm. Sci. 2025,15, 157 18 o 29
To es ima e he c oss-loading, he loading o each indica o should be highe han he
loadings o i s co esponding a iables’ indica o s. Acco ding o Table 5, he c oss-loading
c i e ion is pe ec ; mos o he i ems ha e alues mo e han (0.7) and hei highes alue is
when compa ed wi h o he i ems.
Table 5. C oss-loading esul s.
TR AD SI PE PU BI AT
TR1 0.761 −0.032 0.123 0.042 0.058 0.154 0.077
TR2 0.821 −0.011 −0.075 −0.086 0.220 0.110 −0.048
TR3 0.786 0.109 0.206 0.268 −0.004 0.023 0.068
TR4 0.519 0.197 0.243 0.374 0.056 −0.147 0.237
AD1 0.105 0.657 0.298 0.037 0.143 0.034 −0.045
AD2 −0.009 0.865 0.141 0.121 0.075 0.081 0.140
AD3 0.005 0.845 0.091 0.187 0.202 0.188 0.014
AD4 0.041 0.794 0.055 0.189 0.084 0.244 0.087
SI1 0.039 0.309 0.526 0.011 0.152 0.307 0.233
SI2 0.122 0.093 0.862 −0.022 −0.018 0.023 0.055
SI3 0.090 0.139 0.895 0.133 0.074 0.054 0.039
SI4 0.079 0.125 0.879 0.122 0.046 0.003 0.009
PE1 0.067 0.252 0.109 0.741 0.026 0.050 0.253
PE2 0.148 0.083 0.045 0.689 0.329 0.237 0.092
PE3 0.086 0.172 0.075 0.742 0.307 0.159 0.046
PU1 0.112 0.218 0.025 0.133 0.776 0.120 0.248
PU2 0.103 0.219 0.009 0.253 0.661 0.265 0.303
PU3 0.119 0.119 0.083 0.192 0.817 0.177 0.237
PU4 0.069 0.070 0.106 0.127 0.781 0.241 0.299
BI1 0.061 0.313 0.074 0.221 0.155 0.736 0.216
BI2 0.026 0.282 0.028 0.084 0.320 0.740 0.174
BI3 0.170 0.052 0.087 0.101 0.205 0.785 0.032
AT1 0.033 0.050 0.066 0.139 0.254 0.102 0.827
AT2 0.089 0.032 0.101 0.112 0.204 0.015 0.862
AT3 −0.001 0.078 0.046 0.035 0.253 0.056 0.894
AT4 0.081 0.052 0.008 0.121 0.127 0.216 0.844
No e: The bold alues ep esen he loadings o each indica o on i s co esponding a iable.
4.2. Model Fi Measu es
The model i was assessed h ough eigh indices: CMIN/DF, Roo Mean Squa e E o
o he App oxima ion (RMSEA), Compa a i e Fi Index (CFI), No med Fi Index (NFI),
Tucke -Lewis Index (TLI), Inc emen al Fi Index (IFI), Pa simony- Goodness Measu es
(PGFI), Goodness-o -Fi Index (GFI). The e o e, his s udy con i ms he model i (Table 6)
alidi y acco ding o J. F. Hai e al. (2017).
Table 6. Model i indices.
Measu es o Fi Indices Values Recommended Values
Disc epancy measu emen s CMIN/DF 1.860 (<2)
(RMSEA) 0.065 (0–0.1)
Compa a i e Fi Index (CFI) 0.923 (0.9–1)
Inc emen al adjus men
measu es
No med Fi Index (NFI) 0.902 (0.9–1)
Tucke –Lewis Index (TLI) 0.906 (0.9–1)
Pa simony-adjus ed and
ela ed measu es
Inc emen al Fi Index (IFI) 0.925 (0.9–1)
Pa simony-Goodness Measu es (PGFI) 0.757 (0.5–1)
Goodness-o -Fi Index (GFI) 0.914 (0.9–1)
Adm. Sci. 2025,15, 157 19 o 29
Acco ding o he abo e alues, he o e all i indices indica e ha he model has a good
i o he da a. The CMIN/DF and RMSEA alues all wi hin he accep able ange, and he
inc emen al measu es (CFI, NFI, TLI, IFI), along wi h he pa simony-adjus ed measu es
(PGFI, GFI), all mee o exceed he ecommended h esholds. This comp ehensi e se o
indices con i ms ha he p oposed model is s ong enough and p o ides an adequa e
ep esen a ion o he s udy da a.
4.3. S uc u al Model Assessmen
The s uc u al model was e alua ed by calcula ing he dispa i y be ween dependen
a iables. Acco ding o he o e all model i indices, he indica ion is ha he s uc-
u al model has a good i o he da a. The CMIN/DF = 2.033 and RMSEA = 0.071
con i ms ha hose alues all wi hin he accep able ange and he inc emen al measu es
(
CFI = 0.929
,
NFI = 0.908
, TLI = 0.907, IFI = 0.931), along wi h he pa simony-adjus ed
measu es (
PGFI = 0.787
, GFI = 0.909), all mee o exceed he ecommended h esholds. This
comp ehensi e se o indices con i ms ha he p oposed s uc u al model is s ong enough
and p o ides an adequa e ep esen a ion o he s udy da a. Also, i is es ima ed p ima ily
by pa h coe icien s. A pa h coe icien in a s uc u al model is a numbe ha shows how
wo a iables a e ela ed o one ano he . I shows how he alue o one a iable changes by
one s anda d de ia ion uni when he alue o ano he a iable changes. Typically, pa h
coe icien s all be ween
−
1 and 1, whe e alues nea e
−
1 signi y a s ong nega i e link
and alues nea e 1 indica e a s ong posi i e ela ionship. The e o e, i can clea ly explain
ha he cons uc s o he p oposed model ha e almos s ong posi i e ela ionships.
Acco ding o he pa h analysis (see Figu e 4), Table 7 ound each hypo hesis by
es ima ing he p- alues and he pa h coe icien s. I can be no ed ha six hypo heses a e
suppo ed, while he emaining h ee a e no suppo ed, which in u n indica es ha mos
o he pa hs a e signi ican be ween he independen and dependen a iables.
Adm. Sci. 2025, 15, x FOR PEER REVIEW 19 o 30
Pa simony-adjus ed and
ela ed measu es
Inc emen al Fi Index (IFI) 0.925 (0.9–1)
Pa simony- Goodness Measu es (PGFI) 0.757 (0.5–1)
Goodness-o
-Fi Index (GFI) 0.914 (0.9–1)
Acco ding o he abo e alues, he o e all i indices indica e ha he model has a
good i o he da a. The CMIN/DF and RMSEA alues all wi hin he accep able ange,
and he inc emen al measu es (CFI, NFI, TLI, IFI), along wi h he pa simony-adjus ed
measu es (PGFI, GFI), all mee o exceed he ecommended h esholds. This comp ehen-
si e se o indices con i ms ha he p oposed model is s ong enough and p o ides an
adequa e ep esen a ion o he s udy da a.
4.3. S uc u al Model Assessmen
The s uc u al model was e alua ed by calcula ing he dispa i y be ween dependen
a iables. Acco ding o he o e all model i indices, he indica ion is ha he s uc u al
model has a good i o he da a. The CMIN/DF = 2.033 and RMSEA = 0.071 con i ms ha
hose alues all wi hin he accep able ange and he inc emen al measu es (CFI = 0.929,
NFI = 0.908, TLI = 0.907, IFI = 0.931), along wi h he pa simony-adjus ed measu es (PGFI
= 0.787, GFI = 0.909), all mee o exceed he ecommended h esholds. This comp ehensi e
se o indices con i ms ha he p oposed s uc u al model is s ong enough and p o ides
an adequa e ep esen a ion o he s udy da a. Also, i is es ima ed p ima ily by pa h coe -
icien s. A pa h coefficien in a s uc u al model is a numbe ha shows how wo a iables
a e ela ed o one ano he . I shows how he alue o one a iable changes by one s anda d
de ia ion uni when he alue o ano he a iable changes. Typically, pa h coefficien s all
be ween −1 and 1, whe e alues nea e −1 signi y a s ong nega i e link and alues nea e
1 indica e a s ong posi i e ela ionship. The e o e, i can clea ly explain ha he con-
s uc s o he p oposed model ha e almos s ong posi i e ela ionships.
Acco ding o he pa h analysis (see Figu e 4), Table 7 ound each hypo hesis by es i-
ma ing he p- alues and he pa h coefficien s. I can be no ed ha six hypo heses a e sup-
po ed, while he emaining h ee a e no suppo ed, which in u n indica es ha mos o
he pa hs a e signi ican be ween he independen and dependen a iables.
Figu e 4. Pa h coefficien o s uc u al model.
The esul s o his pape ound ha AT (β = 0.370, CR = 5.481, p < 0.001) has a posi i e
and signi ican impac on BI, sugges ing ha Hypo hesis 1 is suppo ed in his analysis
(see Table 7). I clea ly desc ibes how a i ude has been ound o ha e a signi ican a o -
able impac on he beha io al in en ion o use s who plan o u ilize AI cha bo applica-
ions in he go e nmen sec o in S i Lanka.
Figu e 4. Pa h coe icien o s uc u al model.
The esul s o his pape ound ha AT (
β
= 0.370, CR = 5.481, p< 0.001) has a posi i e
and signi ican impac on BI, sugges ing ha Hypo hesis 1 is suppo ed in his analysis (see
Table 7). I clea ly desc ibes how a i ude has been ound o ha e a signi ican a o able
impac on he beha io al in en ion o use s who plan o u ilize AI cha bo applica ions in
he go e nmen sec o in S i Lanka.
Adm. Sci. 2025,15, 157 20 o 29
Table 7. Hypo heses es esul s.
Hypo hesis Pa h S anda d
Es ima es
S anda d
E o
C i ical
Ra io p-Value
H1 BI ←AT 0.370 0.068 5.481 ***
H2 AT ←PU 0.745 0.127 5.865 ***
H3 AT ←PE −0.041 0.179 −0.228 0.820
H4 PU ←PE 0.855 0.137 6.22 ***
H5 AT ←TR 0.247 0.139 1.777 0.076
H6 PE ←TR 0.401 0.100 4.03 ***
H7 PE ←AD 0.404 0.084 4.829 ***
H8 PU ←SI 0.053 0.094 0.56 0.575
H9 TR ←SI 0.445 0.105 4.217 ***
No e: ***, p< 0.001.
Also, PU (
β
= 0.745, CR = 5.865, p< 0.001) has a posi i e and signi ican impac on
AT, suppo ing Hypo hesis 2. On he o he hand, PE (
β
=
−
0.041, CR =
−
0.228, p= 0.820)
has a nega i e and no sa is ac o y impac on AT, no suppo ing Hypo hesis 3. The h ee
cons uc s o PU, PE, and TR show mixed e ec s on he use a i ude owa d sus ainable
AI cha bo applica ion adop ion in go e nmen adminis a ion se ices. These e ec s a e
ela ed o he use ’s us and belie s. Use a i ude is impo an in in luencing he igh
beha io and ac ion. The h ee cons uc s a ec di e en a i udes. High us will inc ease
wi h ease o use, and ease o use will inc ease use ulness, and hen use ulness will inc ease
use a i ude posi i ely and mo i a e hem o ac posi i ely o use AI cha bo applica ions
in he S i Lanka go e nmen sec o .
PE (
β
= 0.855, CR = 6.22, p< 0.001) has a posi i e and signi ican impac on PU,
suppo ing Hypo hesis 4. I explains he pa h be ween PE and PU. The ela ion be ween
hese wo cons uc s is e y s ong; i is he co e eason o he use ’s a i ude owa d
AI cha bo applica ion adop ion. PE posi i ely in luences he use s’ PU; minimizing he
echnology complexi y inc eases he belie ha AI-based applica ions a e e icien o use
and help ul o go e nmen adminis a ion se ices.
TR (
β
= 0.247, CR = 1.777, p= 0.076) has a posi i e impac on AT, hough his esul is
no s a is ically signi ican , hus no suppo ing Hypo hesis 5. The ela ionship be ween TR
and PE is also posi i e and signi ican (
β
= 0.401, CR = 4.03, p< 0.001), suppo ing Hypo h-
esis 6. This explains he ela ionship be ween pe cei ed ease o use and us . Acco ding o
his s udy, us is a c ucial concep ha in luences o he concep s and in luences he choices
made by consume s. Risk and us a e in e sely co ela ed; as us ises, he es ima ed isk
alls. Building us is he key o boos ing con idence in mode n echnology and i s abili y
o be used mo e e ec i ely wi h less e o . Pe cei ed ease o use is a p ima y componen
o his model ha cha ac e izes he deg ee o complexi y o AI cha bo applica ions o use
in he go e nmen sec o .
Wi h espec o PE, AD (
β
= 0.404, CR = 4.829, p< 0.001) has a posi i e and no
signi ican impac on PE, suppo ing Hypo hesis 7. This ou lines he ela ionship be ween
pe cei ed ease o use and applica ion design and appea ance. The esul s demons a e
ha he applica ion design has a a o able impac on PE o he AI cha bo applica ion; a
well-designed cha bo applica ion enhances use sa is ac ion h ough posi i e in e ac ion
and use. SI (
β
= 0.053, CR = 0.56, p= 0.575) has a posi i e and no sa is ac o y impac on
PU, no suppo ing Hypo hesis 8. Howe e , SI (
β
= 0.445, CR = 4.217, p< 0.001) has a
posi i e and signi ican impac on TR, suppo ing Hypo hesis 9. These esul s desc ibe
he pa h be ween social in luence and PE and us . This pa h shows he posi i e e ec o
social in luence on PU and us , which explains ha social ac o s ha e a ela ionship wi h
mode n echnological applica ions and se ices (posi i e o nega i e) in he go e nmen
Adm. Sci. 2025,15, 157 21 o 29
sec o . The ela ions and communica ions be ween people on social media, in wo k, in he
ma ke s, o any o he place can signi ican ly impac and mo i a e people o us and use
new echnological applica ions in go e nmen sec o o ganiza ions.
4.4. Di ec E ec , Indi ec E ec , and To al E ec
The signi icance o he media ed e ec is accessed by a boo s apping me hod. I was
employed o de i e he di ec , indi ec , and o al e ec s in his model. The esul s (Table 8)
show ha TR indi ec ly in luences PU (
β
= 0.343,
p< 0.01
) and BI (
β
= 0.18,
p< 0.01
),
con i ming i s media ing ole in AI cha bo adop ion. Addi ionally, SI has a s ong e ec
on TR (
β
= 0.445,
p< 0.01
) and indi ec ly in luences PU (
β
= 0.153, p< 0.01),
AT (β= 0.256
,
p< 0.01
), and BI (
β
= 0.095, p< 0.01). Fu he mo e, AD enhances PE (
β
= 0.404,
p< 0.01
)
and PU (
β
= 0.345,
p< 0.01
), which indi ec ly con ibu e o a i ude o ma ion and adop ion
in en ion. These indings highligh ha us , social in luence, and applica ion design
signi ican ly shape he accep ance o AI cha bo s in public adminis a ion se ices in
S i Lanka.
Table 8. Resul s o o al, indi ec , and di ec e ec .
Pa h Es ima es
To al Di ec Indi ec
TR->PE 0.401 0.401 0
TR->PU 0.343 0 0.343
TR->BI 0.180 0 0.180
SI->TR 0.445 0.445 0
SI->PE 0.178 0 0.178
SI->AT 0.256 0 0.256
SI->BI 0.095 0 0.095
AD->PE 0.404 0.404 0
AD->PU 0.345 0 0.345
AD->AT 0.241 0 0.241
AD->BI 0.089 0 0.089
5. Discussion
In ecen yea s, he in eg a ion o a i icial in elligence, pa icula ly cha bo applica-
ions, has su ged wi hin he go e nmen sec o , o e ing inno a i e solu ions o enhancing
ci izen engagemen and s eamlining public se ices. Howe e , he adop ion o hese
echnologies has aced signi ican challenges, as many use s exhibi esis ance o in e ac ing
wi h AI-d i en cha bo s. This pape e alua ed he usabili y o cha bo echnology and
iden i ied he ac o s in luencing use accep ance in a go e nmen adminis a ion se ice
deli e y con ex -based cha bo applica ion in a S i Lankan con ex and designed and al-
ida ed a new esea ch model (ex ended TAM) o success ul AI echnological adop ion
which can be used o u u e ini ia i es in many de eloping coun ies. Gi en he inc easing
p e alence o AI ools and he limi ed exis ing esea ch o guide his inqui y, he de elop-
men o a new model seeks o p o ide esh insigh s and suppo he b oade accep ance o
AI cha bo applica ions. A su ey was conduc ed among a di e se g oup o go e nmen
adminis a ion se ice use s, employing he SEM app oach o analyze he collec ed da a.
In he con ex o social sciences, hese esul s ea i m he necessi y o in eg a ing
echnological ools in bo h he public and p i a e sec o s and adjus ing o he swi ad-
ancemen s in echnology o p omo e he long- e m sus ainable adop ion o AI echnol-
ogy. Howe e , we con end ha de eloping sus ainable AI echnological ini ia i es in he
go e nmen adminis a ion se ice necessi a es mo e han jus implemen ing he newes
echnologies; a he , i calls o a ho ough and calcula ed in eg a ion ha conside s social
Adm. Sci. 2025,15, 157 22 o 29
ac o s, applica ion design and appea ance, and us in he solu ion o challenging issues.
These componen s mus be conside ed by policy p o essionals who can handle oday’s
issues, p o ec he public in e es , and con inue go e nmen o ganiza ions’ se ice deli e y
o he public by e ec i ely u ilizing con empo a y echnological ini ia i es.
The su ey esul s e ealed ha se e al new ac o s signi ican ly in luence use s’
beha io al in en ions owa ds AI cha bo applica ion in go e nmen adminis a ion se ice
deli e y. Acco ding o esea ch (Ma emba & Li,2017), TR ac o can mo i a e indi iduals
o adop echnology. Also, s ong pa ial co ela ions be ween isk and accep ance can be
demons a ed based on esea ch (Eise e al.,2006). Simila ly, he indings o his s udy
highligh ha TR has a s ong di ec impac on PE, while PE di ec ly impac s PU, which
posi i ely impac s use AT and decision making, ul ima ely leading o a change use s’ BIs
o an accep ance o AI cha bo s in go e nmen se ice deli e y. Also, a g oup o esea che s
(Chaouali e al.,2016) ound ha social in luence uly a ec s echnology and e lec s he
deg ee o ai h in i . Simila ly, some s udies disco e ed ha he social en i onmen and
consume decisions signi ican ly in luence use decisions and beha io s (Chaouali e al.,
2016;Fulk e al.,1995;Fulk & Yuan,2017;Malho a & Galle a,1999). Addi ionally, he
esul s o his s udy indica e ha he new ex e nal cons uc , SI, has a s ong impac on
TR and plays a c ucial ole in os e ing us . Responden s la gely exp essed con idence
in hei sa e y and us wo hiness when in e ac ing wi h hese AI-d i en sys ems in
go e nmen sec o in S i Lanka based on he social ecommenda ions. Also, AD posi i ely
in luences PE when ocused on he adop ion o an accep ance o he use o AI cha bo
applica ions in go e nmen sec o o ganiza ions. The e o e, i con i ms he impo ance o
applica ion design/appea ance and conside ing he mul i-na ional language en i onmen
in S i Lankan socie y.
Mo eo e , his esea ch makes a signi ican con ibu ion by demons a ing ha TAM
(F. Da is,1985;F. D. Da is,1989) is su icien o explain how we can ex end he model wi h
ex e nal ac o s such as TR, SI, and AD o iden i y he adop ion o he la es echnology in
di e en en i onmen s. The new esea ch model was buil a ound nine hypo heses and
ocused on analyzing he ela ionships be ween he men ioned ex e nal ac o s and exis ing
TAM co e ac o s. The esul s p o e ha mos o he hypo heses p oposed a e e idence o
he posi i e in luence o us , applica ion design and appea ance, and social in luence on
se e al cons uc s ocused on beha io al in en ion ega ding he accep ance o AI cha bo
applica ions in go e nmen adminis a ion se ices o sus ainabili y.
Howe e , Hypo heses H3, H5, and H8 we e no accep ed acco ding o he esul s o
his s udy. Hypo hesis H3 explains he ela ionship be ween PE o AT. The echnology
accep ance model (F. Da is,1985) sugges s ha PE signi ican ly in luences AT owa ds
echnology. The e o e, we hypo hesized ha i a echnology is pe cei ed as easy o use,
use s would de elop a mo e a o able a i ude owa d adop ing i . The ailu e o p o e his
co ela ion may exhibi he na u e o he go e nmen sec o in S i Lanka, whe e ci izens
migh be less conce ned wi h pe cei ed ease o use due o a highe deg ee o us on
cha bo s (e.g., accessing concu en in o ma ion se ices). Ci izens may p io i ize use ul-
ness and us wo hiness o e pe cei ed ease o use due. Addi ionally, many go e nmen
employees and ci izens may al eady ha e expe ience wi h e-se ices om he p i a e sec o
and, as a esul , he pe cei ed ease o use due ac o migh no be as in luen ial on hei
a i ude as o iginally expec ed in he S i Lankan con ex . Enhancing capabili ies/accu acy
o e ease o use and p omo ing u ili ies/abili ies a he han ease o use can be in oduced
as al e na i e policy measu es o enhance AI cha bo adop ion in S i Lanka’s public sec o .
Simila ly, Hypo hesis H5 explains he ela ionship be ween TR and AT. Acco ding
o p io esea ch s udies, us in echnology accep ance is essen ial, pa icula ly when
ela ed o da a p i acy and secu i y. I was iden i ied as a undamen al de e minan o
Adm. Sci. 2025,15, 157 23 o 29
use s’ a i udes owa ds adop ing new echnologies. Howe e , he esul s o his s udy
in e p e a non-signi ican ela ionship be ween us and a i ude, which could be due o
he speci ic ci cums ances o S i Lanka’s e-go e nmen en i onmen . In a coun y whe e
e-go e nmen se ices a e s ill in he p ocess o being in eg a ed and ea ed, ci izens may
ha e de eloped p ac ical a i udes owa d he new echnology, ocusing mo e on accessibil-
i y and unc ionali y han on us ac o s. The epu a ion and his o y o se ice deli e y o
go e nmen o ganiza ions may also unde es ima e he ole o us in in luencing ci izens’
a i udes owa ds e-go e nmen solu ion accep ance. Also, some al e na i e policy mea-
su es can be aken based on us o enhance AI cha bo adop ion. Enhancing anspa ency
and es ablishing s ong da a secu i y mus be aken wi h hose policy measu es o build
ci izens’ us .
Finally, Hypo hesis H8 desc ibes he ela ionship be ween SI and PU. SI is a widely
ecognized ac o in echnology accep ance models. We hypo hesized ha social in lu-
ence (e.g., ecommenda ions om colleagues, elde s, and leade s) would posi i ely a ec
indi iduals’ pe cep ions o he use ulness o AI cha bo s in go e nmen adminis a ion
se ices. The lack o suppo o his hypo hesis indica es ha SI may ha e a lowe di ec
impac on pe cei ed use ulness in he con ex o go e nmen se ice adop ion in S i Lanka.
This migh happen because pe cei ed use ulness in public sec o echnology adop ion
is mo e likely in luenced by indi idual use expe iences and ask-o ien ed needs a he
han ex e nal social p essu es. F om he pe spec i e o S i Lanka’s e-go e nmen , use s
migh ocus on he p ac ical bene i s o using AI cha bo s such as ease o access and speedy
se ice deli e y a he han how o he s pe cei e hese sys ems. Addi ionally, go e nmen
employees and ci izens migh ha e minimal le els o pe cei ed social in luence o adop
new echnologies such as AI cha bo s, especially i hey a e s ill ela i ely mode n and
no ex ensi ely used so a . Some policy measu es can be aken in his scena io, such as
ocusing on pe o mance/bene i s o e he social popula i y o AI cha bo applica ions.
6. Conclusions
6.1. Resea ch Implica ions
As a de eloping coun y, he go e nmen o S i Lanka and go e nmen policy mak-
e s should conside he esul s achie ed in his s udy o he u u e sus ainabili y in AI
echnological ini ia i es in go e nmen adminis a ion se ices. Ini ially, all AI-based digi-
al go e nmen ini ia i es mus be aligned wi h he long- e m sus ainabili y goals o S i
Lanka. The implemen a ion o mode n echnological mo emen s like AI-based cha bo
applica ions should ake consume s’ us in o conside a ion, suppo ed by go e nmen
egula ion and cus ome expe ience. This means ha he go e nmen mus p io i ize
cus ome us in all digi al go e nmen applica ions. To build up people’s us in digi al
in e ac ions, go e nmen s may p o ide anspa ency by desc ibing how cha bo s ope a e,
wha da a a e ga he ed and p o ec ed, and how use p i acy is main ained. The S i Lankan
go e nmen may implemen s ong cybe secu i y measu es o suppo and egula e he
use o any AI-based applica ions and ini ia e sui able laws and egula ions o con ol he
usage, which will ul ima ely help o build use us . In his scena io, he go e nmen can
o ganize public awa eness p og ams on such new digi al echnologies o emphasize hei
secu i y and eliabili y ega ding sensi i e public da a. Also, egula audi ing me hods,
independen e iews, and public eedback sys ems can be implemen ed o ensu e ha AI
cha bo sys ems emain e hical and e ec i e a p ocessing sensi i e public ela ed da a,
while p ese ing public us .
Secondly, enhancing applica ion design and appea ance can imp o e use amilia i y
and adop ion a es signi ican ly. The go e nmen should make su e ha AI cha bo s ha e
a simple, use iendly design wi h com o able na iga ion. As desc ibed p e iously, o
Adm. Sci. 2025,15, 157 24 o 29
accommoda e a ied popula ions, accessibili y ea u es such as mul ilingual suppo in
oice assis ance and impo an possibili ies o disabled use s mus be in eg a ed. Also,
consis ency among di e en pla o ms should be main ained, especially on mobile applica-
ions, go e nmen websi es and in o ma ion po als, and social media ne wo ks, o suppo
highe usabili y and accep abili y o AI cha bo sys ems.
Finally, based on he ou comes o social in luence, he go e nmen may ake s eps o
p omo e AI cha bo s ac i ely h ough digi al awa eness p omo ions. Popula in luence s
and well-known go e nmen policy make s and o ice s can make endo semen s on go e n-
men AI solu ions o build and enhance social con idence ul ima ely. P esen ing eal-wo ld
accomplishmen s can boos c edibili y and inspi e wide AI solu ion adop ion. Public
engagemen measu es, such as dynamic cha bo exhibi ions and incen i es o encou age
ea ly adop e s, can help o inc ease use in e es . Go e nmen s can speed up he ansi ion
o AI-powe ed public se ice engagemen s by cul i a ing a cul u e ha no malizes and
encou ages cha bo use in e-go e nmen se ices.
These ecommenda ions can help guide policymake s and applica ion designe s in
de eloping a s ong, p o ec ed, and us wo hy amewo k o AI adop ion in go e nmen
se ices, suppo ing he sus ainabili y o digi al go e nmen ini ia i es in S i Lanka.
6.2. Limi a ions and Fu u e Resea ch
This s udy is only designed o measu e he in en ion o use s, no he ac ual usage a e
o such mode n echnology. This p o ides di ec ion o u u e esea ch o use ac ual usage
da a o suppo ou indings. To s eng hen he alidi y o his s udy’s indings, u u e
esea ch should mo e beyond use pe cep ion su ey da a and inco po a e eal-wo ld
adop ion me ics. In ha scena io, u u e esea che s should moni o he usage in o ma ion
o AI cha bo s in go e nmen po als o e a easonable ime du a ion (p e e ably six mon hs
o one yea ). Then, p e-adop ion expec a ions (su ey da a on use expec a ions) should be
compa ed wi h ac ual pos -adop ion beha io s (ac ual usage s a is ics) o ob ain alida ed
conclusions. Also, i is ecommended o check he usage pa e ns o di e en AI cha bo
e sions by di e en solu ion in e aces. This can be help ul o check he use beha io
based on applica ion design implica ions. Fu he mo e, i is impo an o no e ha he
esul s o he s udy can be imp o ed u he by inc easing he sample size o he use s o
he go e nmen adminis a ion se ices in S i Lanka.
Au ho Con ibu ions: Concep ualiza ion, Y.A.; me hodology, A.S.R.; so wa e, A.S.R.; alida ion,
Y.A. and T.D.H.N.N.; o mal analysis, A.S.R.; in es iga ion, A.S.R.; esou ces, A.S.R.; da a cu a ion,
A.S.R.; w i ing—o iginal d a p epa a ion, A.S.R.; w i ing— e iew and edi ing, T.D.H.N.N.; isual-
iza ion, A.S.R.; supe ision, Y.A; p ojec adminis a ion, Y.A.; unding acquisi ion, Y.A. All au ho s
ha e ead and ag eed o he published e sion o he manusc ip .
Funding: This esea ch ecei ed no ex e nal unding.
Ins i u ional Re iew Boa d S a emen : E hical e iew and app o al we e wai ed o his s udy due
o Hanyang Uni e si y Resea ch E hics Policy exemp ions (h ps://i b.hanyang.ac.k /02
_
guide/
guide02.h ml, accessed on 13 Ma ch 2025).
In o med Consen S a emen : In o med consen was ob ained om all subjec s in ol ed in he s udy.
Da a A ailabili y S a emen : The da a p esen ed in his s udy a e a ailable on eques om he
co esponding au ho due o p i acy and e hical easons.
Con lic s o In e es : The au ho s decla e no con lic o in e es .
