Yue, Yao; Ng, Siew Imm; No azlyn Kamal Basha
A icle
Consump ion alues, a i udes and con inuance in en ion
o adop Cha GPT-d i en e-Comme ce AI Cha bo
(LazzieCha )
Pakis an Jou nal o Comme ce and Social Sciences (PJCSS)
P o ided in Coope a ion wi h:
Joha Educa ion Socie y, Pakis an (JESPK)
Sugges ed Ci a ion: Yue, Yao; Ng, Siew Imm; No azlyn Kamal Basha (2024) : Consump ion alues,
a i udes and con inuance in en ion o adop Cha GPT-d i en e-Comme ce AI Cha bo (LazzieCha ),
Pakis an Jou nal o Comme ce and Social Sciences (PJCSS), ISSN 2309-8619, Joha Educa ion Socie y,
Pakis an (JESPK), Laho e, Vol. 18, Iss. 2, pp. 249-284
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/301668
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/4.0/
Pakis an Jou nal o Comme ce and Social Sciences
2024, Vol. 18(2), 249-284
Pak J Comme Soc Sci
Consump ion Values, A i udes and Con inuance
In en ion o Adop Cha GPT-d i en E-Comme ce AI
Cha bo (LazzieCha )
YAO YUE (Co esponding au ho )
School o Business and Economics, Uni e si i Pu a Malaysia, Selango , Malaysia
Email: gs60[email p o ec ed]pm.edu.my
Siew-Imm Ng
School o Business and Economics, Uni e si i Pu a Malaysia, Selango , Malaysia
Email: imm_[email p o ec ed]u.my
No azlyn Kamal Basha
School o Business and Economics, Uni e si i Pu a Malaysia, Selango , Malaysia
Email: no[email p o ec ed]u.my
A icle His o y
Recei ed: 01 Ma 2024
Re ised: 11 June 2024
Accep ed: 19 June 2024
Published: 30 June 2024
Abs ac
The apid ise o cha gene a i e p e- ained ans o me (Cha GPT) has b ough huge
oppo uni ies o e-comme ce pla o ms o use i o consume communica ion and se ice.
This pape p oposes a esea ch amewo k o expand he alue-a i ude-beha io model by
in eg a ing he heo y o consump ion alues. The pu pose is o explo e key consump ion
alues ha a ec consume s' a i udes and hei con inuance in en ion o adop Cha GPT-
d i en cha bo s (i.e., LazzieCha ) in he e-comme ce se ices con ex while examining he
mode a ing e ec o online shopping sel -e icacy in his p ocess. A ques ionnai e su ey
was conduc ed in Singapo e, in ol ing 305 pa icipan s who had expe ienced LazzieCha
in Lazada. Use he s uc u al equa ion modelling o Sma s 4 o analyze he da a. The
esul s show ha consume 'a i udes owa ds LazzieCha ha e signi ican ly a ec ed hei
con inuance in en ion o adop LazzieCha , and online shopping sel -e icacy has
mode a ed his ela ionship. In addi ion, he esea ch con i ms ha emo ional alue and
epis emic alue a e he main d i ing ac o s o consume 'a i udes, ollowed by unc ional
alue, social alue and epis emic alue. Consume s' a i udes owa ds LazzieCha media e
he ela ionship be ween unc ional alue, social alue, epis emic alue, and con inuance
in en ion o adop LazzieCha . By de eloping a esea ch amewo k ha ollows he logic
o alue-a i ude-beha io model, his s udy p o ides heo e ical and p ac ical insigh s
based on he heo y o consume alues, which could guide e-comme ce pla o ms o mo e
Cha GPT-d i en E-Comme ce AI Cha bo (LazzieCha )
250
e ec i ely ca e o hei consume s' con inuance in en ion o adop ing Cha GPT-d i en e-
comme ce AI cha bo se ice.
Keywo ds: Theo y o consump ion alue, he alue-a i ude-beha io (VAB) model,
Cha GPT, cha bo s, LazzieCha con inuance in en ion, inline shopping sel -e icacy.
1. In oduc ion
The apid de elopmen o a i icial in elligence (AI) has changed he online consume
se ice expe ience and c ea ed g ea oppo uni ies o companies o in e ac wi h hei
consume s ia cha bo s. (Chen e al., 2021; Hollebeek e al., 2023; Kuma e al., 2019;
Sidaoui e al., 2020). In Singapo e, nea ly hal o IT p o essionals ha e epo ed accele a ed
AI ool de elopmen , pa icula ly cha bo s (S ai s Times, 2021). E-comme ce leads in
cha bo adop ion, wi h p edic ions o o e $100 billion in ansac ions by 2023
(Se iceBell, 2023). This end e lec s he inc easing use o AI-d i en cha bo s o enhance
e-comme ce consume se ice (Chen e al., 2021; Sidaoui e al., 2020).
Al hough he adop ion a e o cha bo s is ge ing highe and highe in e-comme ce
consume se ice, he e a e challenges like de elopmen expe ise and con ex awa eness
(Lei e al., 2021), leading o issues like epe i i e esponses and delayed eplies (iiMedia
Resea ch, 2021). To add ess hese challenges, ad anced AI ools like Cha Gene a i e P e-
T ained T ans o me (Cha GPT) o e p ecise and pe sonalized esponses
(GlobeNewswi e, 2023). Since i s launch on No embe 30, 2022, Cha GPT, a new
gene a ion o gene a i e a i icial in elligence cha bo , has b ough su p ises o use s wi h
i s e olu iona y unc ions, which can answe con e sa ion allies and show s onge
p o essional knowledge and awa eness o con ex (Ande s, 2023). Hence, he Cha GPT-
d i en cha bo is highly adap able and has al eady been in eg a ed in o a ious e-comme ce
pla o ms (Jumpselle , 2023), such as K onos, Ai India, Expedia, and eBay
(GlobeNewswi e, 2023; Ou lookindia, 2023; In e ne Re ailing, 2023). Besides, e-
comme ce gian s such as Alibaba, Amazon and Lazada ha e implemen ed Gene a i e
a i icial in elligence cha bo applica ions o enhance hei e-comme ce ope a ions
capabili ies (Kshe i, 2024; Yo dan, 2023). Fo example, Lazada consume s can access
LazzieCha by swiping down he Lazada app homepage (Lazada, 2023). LazzieCha can
na u ally answe i s consume s' shopping que ies and ecommend ele an p oduc s o
hemes o consume s, and can also ind p oduc desc ip ions and link hem o p oduc s
p o ided on he e-comme ce pla o m (Yo dan, 2023), which p omo es LazzieCha o
achie e an imp essi e u iliza ion a e wi h he suppo o Cha GPT-d i en e-comme ce AI
cha obo (Lazada, 2023; Simila Web, 2023). This shows ha using Cha GPT as a cha bo
o online consume expe ience in e-comme ce is a g owing end wi h g ea ma ke
po en ial. Howe e , he e is limi ed esea ch ha ocuses on consume s' con inuance
in en ion o adop Cha GPT-d i en cha bo s, speci ically LazzieCha .
Yue, Ng & Basha
251
Academic s udies ha e explo ed AI cha bo s in e-comme ce, emphasizing usabili y,
empa hy, pe sonali y, and consume a i udes (Hollebeek e al., 2023; Kuma e al., 2019;
Sidaoui e al., 2022). In he Cha GPT-d i en con ex , Kshe i (2024) sugges ed ha
Cha GPT-d i en cha bo can p o ide consume s wi h a pe sonalized and in e ac i e e-
comme ce expe ience in e-comme ce, which g ea ly imp o es e iciency and s eng hens
alue p oposi ion. Ne e heless, empi ical da a on Cha GPT-d i en cha bo s in e-
comme ce a e sca ce, and he ole o consume pe cei ed alues is unde explo ed (Hu e
al., 2018; Kshe i, 2024)). Hence, he e is s ill a gap in he dimension o alue p oposi ion
conce ning he online consume expe ience wi h Cha GPT-d i en cha bo adop ion in e-
comme ce, while he mode a ing e ec s on cha bo adop ion emain unde explo ed (Chen
e al., 2023). Meanwhile, S zelecki (2023) called o mo e li e a u e abou indi iduals’
in en ions o adop Cha GPT. The e o e, his pape a emp s o use he Cha GPT-d i en e-
comme ce cha bo (i.e., LazzieCha ) o u he esea ch in his ield. In o he wo ds, his
s udy ex ends he consump ion alues heo y and ollows alue-a i ude-beha io logic o
in es iga e ac o s in luencing consume a i udes and con inuance in en ion o adop
LazzieCha , conside ing he mode a ing e ec o online shopping sel -e icacy in his
p ocess.
In o he wo ds, his s udy aims o add ess he esea ch gap by p o iding empi ical da a on
consume s' con inuance in en ion o adop Cha GPT-d i en e-comme ce AI cha bo s (i.e.,
LazzieCha ). By explo ing he ole o consump ion alues and he mode a ing e ec o
online shopping sel -e icacy, he s udy con ibu es o a be e unde s anding o consume
alue p oposi ion and decision-making p ocesses ega ding he adop ion o ad anced
Cha GPT-d i en cha bo s. The indings o his s udy a e o p ac ical signi icance o e-
comme ce en e p ises and cha bo de elope s. By unde s anding he ac o s ha a ec
consume s ' alue p oposi ion, a i udes and in en ions owa ds Cha GPT-d i en cha bo s,
companies can cus omize hei consume se ice s a egies and implemen a ion o cha bo s
o be e mee consume s' needs and p e e ences. Addi ionally, he s udy's heo e ical
con ibu ions can in o m u u e esea ch in he ield o consume beha io and he adop ion
o AI-powe ed echnologies in e-comme ce and beyond.
This pape add esses h ee key esea ch ques ions: (1) Do consump ion alues in luence
a i udes owa ds LazzieCha ? (2) Do a i udes media e he ela ionship be ween
consump ion alues and he in en ion o adop LazzieCha ? (3) Does online shopping sel -
e icacy mode a e he link be ween a i udes and he in en ion o adop LazzieCha ?
The pape p oceeds wi h a li e a u e e iew, concep ual model, and hypo hesis
de elopmen , ollowed by me hods and esul s, and concludes wi h discussions on
con ibu ions, implica ions, limi a ions, and u u e esea ch di ec ions.
Cha GPT-d i en E-Comme ce AI Cha bo (LazzieCha )
252
2. Li e a u e Re iew
2.1. Cha bo and Cha GPT-d i en Cha bo
P e ious s udies ha e p oposed a ious de ini ions o cha bo s. Fo example, hey a e
desc ibed as so wa e agen s ha p omo e au oma ic dialogue h ough na u al language
p ocessing" (B and zaeg and Fols ad, 2017), a i icial s uc u es o human in e ac ion using
na u al language as inpu and ou pu " (B ennan, 2006), and "dialogue en i ies o compu e
p og ams d i en by a i icial in elligence espond o use s' que ies h ough ex -based
dialogue" (Shawa and A well, 2007). In his s udy, he cha bo was de ined as a ool ha
deli e s immedia e esponses o eques s ia ex -based in e ac ions, ans o ming
business-consume communica ion and in e ac ion wi hou empo al o spa ial es ic ions
and elimina ing he necessi y o human in ol emen (Ga ziou a and Sap ikis, 2022).
Meanwhile, Cha GPT, a gene a ion p e- aining con e e cha bo applica ion de eloped
by OpenAI, is good a gene a ing na u al language ex o sol e complex p oblems ha
equi e ad anced analysis (Na u al e al., 2023). I gene a es human-like esponses,
mimicking s a is ical language pa e ns ound online, e ec i ely add essing a ious use
que ies (Gilson e al. 2022; Fe nandez, 2023; Haleem e al., 2023; Lund and Wang, 2023).
Unlike adi ional cha bo s ha ely on p ede ined ules limi ed o speci ic domains,
Cha GPT uses machine lea ning o gene a e esponses om a la ge amoun o ex da a
(Panda and Kau , 2023). This allows add essing b oade que ies, including no el ones
adi ional cha bo s s uggle wi h, by con inuously lea ning om new da a wi hou manual
ule upda es.
Cha bo s a e widely in eg a ed in o business s a egies and consume se ice, o e ing 24/7
na u al language capabili ies mimicking human con e sa ion (Zums ein and Hunde ma k,
2018). E ec i e cha bo in eg a ion p o ides high-quali y, as se ice (Co de o e al.,
2022). In e-comme ce, cha bo s le e age da a o add ess consume inqui ies, enhancing he
shopping expe ience wi h con enien , pe sonalized in e ac ions o build consume
ela ionships and educe unce ain y (Cui e al., 2017; Chen e al., 2021). Haleem e al.
(2023) highligh Cha GPT's powe in e alua ing e ail s a egies, consume bases, and
p oduc a ibu es o iden i y success easons and ecommend enhancemen s. Kuma e al.
(2023) deem Cha GPT indispensable o e aile s, deli e ing seamless consume
expe iences h ough pe sonalized ecommenda ions, e icien inqui y handling, and ound-
he-clock assis ance. Hence, in eg a ing Cha GPT-d i en cha bo s in o e-comme ce
bene i s businesses and consume s (Kshe i, 2024).
2.2 Value–A i ude–Beha io Model
Ga ziou a and Sap ikis (2022) conduc ed a comp ehensi e li e a u e e iew and ound ha
esea ch on consume se ice cha bo s is a highly ele an and popula opic a p esen , and
key ac o s a ec ing he in en ion o adop ing cha bo s include consume s 'pe o mance
expec a ions and a i udes. Pasupule i and Thiyyagu a (2024) also ound ha a posi i e
a i ude owa ds Cha GPT is mo e likely o con inue o adop i and ecommend i o o he s.
Yue, Ng & Basha
253
Hence, a i udes may di ec ly a ec ac o consume 's con inuance in en ions o adop
Cha GPT-d i en e-comme ce cha bo . Besides, Ga ziou a and Sap ikis (2022) also
emphasize he impo ance o u he explo a ion o iden i y addi ional ac o s and
dimensions a ec ing cha bo adop ion in he e-comme ce domain. To add ess his call in
he li e a u e, his pape explo es he logical ela ionship among consump ion alues,
a i udes and con inuance in en ions o adop LazzieCha .
Fu he mo e, exis ing li e a u e has explo ed he in e play be ween alues, a i udes, and
beha io al in en ions in online o in o ma ion echnology con ex s h ough a ious
pe spec i es (Hasan, 2022). Fo ins ance, Jayawa dhena (2004) highligh ed ha
consume s' alues, encompassing ai s like sel -di ec ion, hedonism, and achie emen ,
exe a posi i e in luence on hei a i udes owa d e-shopping. These a i udes, in u n,
shape hei in en ions o adop e-comme ce, wi h a i udes se ing as in e media ies
be ween alues and beha io . Kang (2014) demons a ed ha mone a y, con enience,
emo ional, and social alues indi ec ly impac consume s' con inued adop ion o
augmen ed eali y and mo ion cap u e appa el in e-comme ce h ough he media ing ole
o a i udes, pa icula ly u ili a ian pe o mance expec a ions. Xu e al. (2020) ound ha
consume s' alue pe cep ions ega ding echnology usage in es au an s a e in luenced by
a i udes, encompassing aspec s like pleasu e and u ili a ian expec a ions, which
subsequen ly d i e consume s o adop onsi e es au an in e ac i e sel -se ice
echnology. Addi ionally, acco ding o Dwi edi e al. (2019), a i udes play a c ucial ole
in he accep ance and u iliza ion o inno a i e echnologies and applica ions. Yan e al.
(2021) emphasized ha a i udes a e c i ical an eceden s in he con inued adop ion o online
echnologies.
Value-A i ude-Beha io (VAB) model is a sequen ial p og ession om alues o a i udes
o beha io is a obus amewo k in esea ch and has gained widesp ead accep ance as a
amewo k o assessing measu emen and s uc u al models aimed a explaining o
p edic ing beha io al in en ions, wi h a i udes se ing as a media ing ac o ha connec s
indi idual alues and beha io s (Home and Kahle, 1988; Johns on e al., 2023). The e o e,
employing he VAB model logically no only enhances ou comp ehension and explo a ion
o he undamen al d i ing ac o s and cha ac e is ics unde lying consume s' con inued
adop ion o LazzieCha bu also di ec s he a en ion o schola s and p ac i ione s owa d
he key ole o a i udes as a c ucial b idge acili a ing he con inued adop ion o Cha GPT-
d i en cha bo s.
2.3 Theo y o Consump ion Value
Bo h ma ke ing and in o ma ion sys ems esea che s ha e long held a keen in e es in he
concep o alue (Ka jaluo o e al. 2021). When i comes o accep ing p oduc s and
se ices, consump ion alue as a key de e minan o consume decision-making has been
widely s udied (Ka jaluo o e al. 2021; She h e al., 1991; Tu el e al. 2010). P e ious
s udies ha e also e ec i ely applied he heo y o consume alue o ca e ully s udy he
Cha GPT-d i en E-Comme ce AI Cha bo (LazzieCha )
254
con inuous beha io o use s in he ne wo k en i onmen , including mobile echnology
(Yang and Lin, 2017) and augmen ed eali y applica ions (Wang e al. 2022), and mobile
banking se ices (Ka jaluo o e al. 2021). Hence, he a ionale behind consume adop ion
se ice decision-making can be explained ia he heo y o consume alue (She h e al.,
1991). This heo y o e s a obus amewo k o p edic ing, elucida ing, o cha ac e izing
consume -le el choices ega ding p oduc o se ice adop ion by delinea ing i e dis inc
consume alues: unc ional, social, emo ional, epis emic, and condi ional (She h e al.,
1991).
Pas li e a u e esea ch demons a ed he e ec i eness o he heo y o consump ion alue
in bo h concep ualizing and in es iga ing he adop ion o digi al applica ions and se ices.
Fo example, Tu el e al. (2010) con end ha he heo y o consump ion alue can explain
why use s op o adop digi al p oduc s and se ices a e e alua ing mul iple ace s o
alue. Kau e al. (2018) disclosed ha sus ained usage o online social media b and
communi ies is in luenced, in pa , by social and emo ional alues. Fu he mo e, Kau e
al. (2021) iden i ied epis emic alue as he p ima y d i ing ac o behind consume
adop ion o ood deli e y apps, ollowed by condi ional, unc ional, and social alue.
Omigie e al. (2017) poin ed ou ha consume s' con inued adop ion o M-PESA mobile
inancial se ices is posi i ely a ec ed by consump ion alues like aes he ics,
con enience, mone a y aspec s, epis emic ac o s, and sel -g a i ica ion. The e o e, in he
online en i onmen , we should ei e a e he heo y o consump ion alue ele ance in
p edic ing speci ic alue p oposi ions, which can help us o unde s and and guide
consume s o con inue o adop Cha GPT-d i en cha bo s (i.e., LazzieCha ) in e-
comme ce. Besides, Kau e al. (2018) ha e obse ed ha condi ional alue p edominan ly
add esses consume decision-making wi hin a ma ke ing con ex and should be excluded
om conside a ion when i comes o decisions ela ed o he adop ion o in o ma ion
echnology. Consequen ly, his pape concen a es on unc ional alue, social alue,
emo ional alue, and epis emic alue.
3. Resea ch Hypo heses
3.1 Func ional Value
Func ional alue is ela ed o consume s' pe cei ed bene i s om he ac ual and physical u ili y
o a p oduc o se ice, including i s cha ac e is ics, a ibu es o p ac icali y (She h e al. 1991).
The unc ional alue o any p oduc o se ice can be gauged by he u ili a ian ad an ages
consume s gain om i s angible pe o mance (Suki, 2013). Consume s' choices o adop a
p oduc o se ice is a ec ed by i s impo an unc ional cha ac e is ics and he alue gene a ed
by i s p ac ical o physical a ibu es. In he con ex o mobile applica ions, consume s' g ea e
pe cei ed unc ional alue is ela ed o he highe possibili y o o ming a good a i ude and
ecei ing se ice (Goh e al. 2014). Ea ly s udies ha e de e mined he unc ional alue o online
e ail op ions, including e iciency, economic alue, ime-sa ing, ask comple ion and
success ul esul s (And ews, 2007).
Yue, Ng & Basha
255
The e a e se e al s udies emphasize he key ole o he unc ional cha ac e is ics o cha bo s in
shaping consume s' in en ions and a i udes. Fo example, Cicco e al. (2020) emphasize ha
adop ing cha bo s enables e aile s o mee he needs o young consume s any ime and
anywhe e. Cha bo s' unc ionali y is mainly e lec ed in he in e ac ion be ween consume s and
cha bo s, hus a ec ing hei a i udes owa ds e ail b ands (Za ouali e al. 2018) and B and
p e e ences (T i edi, 2019). The e o e, he unc ional alue o cha bo s is ela ed o he se ices
and unc ions hey p o ide, such as answe ing use s' que ies, p o iding high-quali y ela ed
in o ma ion, and showing high esponsi eness, which minimizes he ene gy and ime spen by
use s o a speci ic pu pose (Gilson e al. 2022; Fe nandez, 2023). As he la es e sion o
cha bo , he Cha GPT-d i en e-comme ce cha bo has he po en ial o o e mo e unc ional
alue (e.g., s onge p o essional knowledge and awa eness o con ex ) han i s p edecesso s,
po en ially leading o mo e posi i e use a i udes owa d i s adop ion. In o he wo ds, he
unc ional ad an age o Cha GPT is ha i can unde s and he backg ound o he dialogue o
gene a e mo e ealis ic, cohe en and pe sonalized esponses. Besides, Cha GPT p o ides many
unc ional ad an ages, including p o iding de ailed p oduc desc ip ions, p omo ing i ual y-
on expe iences and he abili y o seamlessly in eg a e wi h augmen ed o i ual eali y
echnologies. These capabili ies empowe consume s wi h up- o-da e knowledge and high-
quali y, aluable unc ionali ies (Kuma e al. 2023).
In he ield o e-comme ce, LazzieCha is he la es Cha GPT-d i en cha obo in eg a ed in o
he Lazada pla o m. LazzieCha is good a unde s anding hei consume s' que ies quickly and
esponding na u ally and in ui i ely. In addi ion, i can ecommend ela ed p oduc s o hemes
acco ding o consume s 'in e es s. Consume can easily access p oduc desc ip ions in he cha
in e ace and di ec ly na iga e o Lazada's p oduc s. This simpli ied and e icien p ocess
ensu es ha consume s a e well-in o med and can make a pu chase quickly, con enien ly and
con iden ly. In o he wo ds, unc ional alue plays an impo an ole in shaping consume s
'a i udes owa ds LazzieCha . The li e a u e emphasizes he impo ance o unc ional
cha ac e is ics, such as e iciency, ime sa ing and success ul esul s, which in luence he
adop ion and a i udes o consume s. The e o e, his s udy posi s ha :
➢ H1. Func ional alue is posi i ely ela ed o he consume 's a i udes owa ds
LazzieCha .
3.2 Social Value
Social alue was desc ibed as he pe cei ed bene i de i ed om he connec ion be ween an
op ion and one o mo e speci ic social g oups by She h e al. (1991), which is a ained h ough
he use o obse able p oduc s o se ices and is in luenced by he social in luence o e e ence
g oups (She h e al., 1991). Consume beha io is o en d i en by he desi e o social s a us
(Holb ook, 1999), which can also be a p ima y mo i a o o he adop ion o p oduc s and
se ices (Roge s, 1995). In he ealm o digi al echnology, social alue encompasses he
emo ional g a i ica ion, boos in sel -es eem, and sense o belonging ha use s expe ience
h ough he use o applica ions (Jo dan, 2008). Wu e al. (2018) ound ha a i udes owa ds
Cha GPT-d i en E-Comme ce AI Cha bo (LazzieCha )
256
hedonism and u ili a ianism can egula e he ela ionship be ween social alue and pu chase
in en ion in he online social shopping en i onmen . Kau e al. (2021) and Omigie e al. (2017)
epo ed ha social alues ha e a posi i e impac on use s' in en ion o adop digi al echnology
se ices.
On he o he hand, Sweeney and Sou a (2001) ound ha social alue encompasses social
accep ance, in e pe sonal ela ionships, and he social image associa ed wi h he se ice being
accep ed by consume s. When consume s pe cei e digi al echnology se ices as a means o
enhance hei sel -image, exp ess hei pe sonali y, o posi ion hemsel es wi hin a speci ic
social s a us, hei pe cep ion o social alue d i es hem o exhibi a s ong a i ude and
in en ion o adop his se ice (Chaouali e al., 2023). The e o e, in his s udy, consume social
alue can be pe cei ed in he con ex o iends, colleagues, o amily membe s using he
Cha GPT-d i en cha bo se ice (i.e., LazzieCha ). This se ice o e s pe sonalized ma ke ing
in o ma ion and ad ice based on consume s' p e ious beha io s (such as online commen s,
e iews, and social media pos s), needs, p e e ences, and consump ion pa e ns, he eby
enhancing consume s' emo ional sa is ac ion and hei pe cep ion o social s a us (Ca alho and
I ano , 2023). LazzieCha p o ides g ea e social alue (e.g., in e ms o image and social
ela ionships) compa ed o adi ional cha bo se ices, hus esul ing in consume s' posi i e
a i udes owa d he adop ion o LazzieCha in consume se ice. The e o e, his s udy posi s
ha :
➢ H2. Social alue is posi i ely ela ed o consume s' a i udes owa d LazzieCha .
3.3 Emo ional Value
Emo ional alue can be de ined as he pe cei ed bene i ha consume s de i e om he abili y
o e oke sensa ions, memo ies, emo ional s a es, and he impac o emo ions (She h e al., 1991).
The pas li e a u e emphasizes he impo ance o emo ional alue in in luencing consume s
'a i udes and beha io s owa ds digi al echnology applica ions (including a i icial in elligence
cha bo s). In o he wo ds, emo ional alue pe ains o he posi i e and enjoyable eelings ha
a ise du ing he use o digi al echnology applica ions (Be aies e al., 2017). Emo ional alue
has he capaci y o igge consume s' emo ional s a es, consequen ly in luencing hei
inclina ions o beha iou (Yang e al., 2015). Peng e al. (2014) epo ed ha he p esence o
emo ional alue wi hin he domain o mobile digi al echnology applica ions and i s e ec on
consume s' a i udes owa d b ands. Ka jaluo o e al. (2021) p o ided e idence ha emo ional
alue s ands as a key a iable d i ing o ce behind consume s' in en ions o adop new digi al
echnology applica ions. P e ious s udies ha e also indica ed ha AI cha bo s can be in ol ed
in simple con e sa ions and e ec i ely o e solu ions o consume s' que ies, in luencing
consume s' emo ions o a i udes wi hin he e-comme ce con ex (Kasilingam, 2020). In o he
wo ds, he abili y o AI cha bo s o p o ide accu a e ad ice ins ils a sense o being unde s ood
by consume s, making hem eel ha he AI cha bo genuinely ca es abou hei emo ions (Chen
e al., 2023). LazzieCha , as he newes Cha GPT-d i en cha bo in eg a ion, aims o ele a e
consume se ice by deli e ing supe io se ice and emo ional expe iences compa ed o
p e ious AI cha bo s. LazzieCha can o e high-quali y consume se ice and e ec i e p oduc
Yue, Ng & Basha
263
2003; Hai e al., 2013). In o he wo ds, he basic p inciple o selec ing PLS-SEM is based
on he non-no mal dis ibu ion o collec ed da a, among which PLS-SEM is conside ed o
be mo e sui able o analyzing non-no mal dis ibu ion da a (Hai e al., 2017). Inspi ed by
he abo e esea ch, he au ho uses PLS-SEM echnology o analyze he esea ch esul s.
Meanwhile, he componen -based me hod was adop ed by he boo s apping echnology o
s ic ly e alua e all he p oposed assump ions h ough a la ge numbe o i e a ions (n =
5,000). Besides, since da a was collec ed om a single sou ce, we i s es ed he common
me hod bias acco ding o he ecommenda ions o Kock and Lynn (2012) and Kock (2015).
In his me hod, all he a iables will be eg essed acco ding o a common a iable. I he
VIF alue is less han o equal o 3.3, he e is no de ia ion in single-sou ce da a. The
analysis esul (see Table 3) shows all a iable VIF alues less han 3.3 hus single sou ce
bias is no a se ious issue wi h ou da a.
Table 3: Full Collinea i y Tes ing
FUV
SOV
EMV
EPV
ATL
OSS
1.265
1.232
1.398
1.362
1.307
1.291
No e (s): FUV, Func ional Value; SOV, Social Value; EMV, Emo ional Value; EPV, Epis emic Value;
ATL, A i udes owa d LazzieCha ; OSS, Online shopping Sel -e icacy
4.4 Measu emen Model
We use a ious indica o s o e alua e he model using a ious me ics, which co e
loadings, composi e eliabili y (CR), and a e age a iance ex ac ed (AVE). To be speci ic,
ac o loadings alue > 0.5 means a e indica i e o good measu emen i em consis ency
(Chin, 1998). Nunnally and Be ns ein (1994) s udied ha he alue o composi e eliabili y
(CR) be ween 0.7 and 0.9 is conside ed sa is ac o y. Besides, C onbach's alpha alue is
usually ega ded o be eliable da a a alues g ea e han 0.7 (Ne emeye e al., 2003).
Fo nell and La cke (1981) sugges ed ha he A e age Va iance Ex ac ed (AVE) should
exceed he minimum cu o alue o 0.5 o ob ain su icien con e gence e ec i eness.
Table 4 shows his s udy's analysis esul o he cons uc alidi y o he model. The ac o
loading alues ange om 0.868 o 0.668, highe han 0.5. C onbach's alpha alues ange
om 0.812 o 0.894, highe han 0.7. The AVE alues ange om 0.585 o 0.631, highe
han 0.5. The CR alues om 0.813 o 0.896 which a e all abo e 0.7.
Cha GPT-d i en E-Comme ce AI Cha bo (LazzieCha )
264
Table 4: Loadings, Reliabili y and Validi y
Cons uc s
I ems
Mean
(SD)
Ou e
VIF
Fac o
Loadings
C onbach’s
alpha
CR
AVE
Func ional
Value
0.849
0.853
0.585
FUV1
3.357
(1.104)
1.934
0.809
FUV2
3.311
(1.127)
1.943
0.779
FUV3
3.361
(1.154)
1.861
0.794
FUV4
3.407
(1.162)
1.923
0.668
Social Value
0.812
0.813
0.591
SOV1
3.289
(1.152)
1.726
0.782
SOV2
3.262
(1.100)
1.837
0.741
SOV3
3.321
(1.149)
1.753
0.782
Emo ional
Value
0.862
0.865
0.611
EMV1
3.426
(1.135)
1.862
0.705
EMV2
3.485
(1.134)
2.028
0.815
EMV3
3.528
(1.173)
2.118
0.809
EMV4
3.548
(1.150)
2.173
0.793
Epis emic
Value
0.836
0.840
0.631
EPV1
3.459
(1.142)
1.988
0.737
EPV2
3.407
(1.139)
1.913
0.782
EPV3
3.430
(1.169)
1.946
0.858
A i udes
Towa d
LazzieCha
0.825
0.825
0.611
ATL1
3.328
(1.238)
1.757
0.801
ATL2
3.308
(1.186)
1.932
0.750
ATL3
3.367
(1.194)
1.934
0.791
Online
Shopping Sel -
e icacy
0.894
0.896
0.627
OSS1
3.387
(1.221)
2.189
0.865
OSS2
3.426
2.111
0.816
Yue, Ng & Basha
265
(1.188)
OSS3
3.433
(1.205)
2.299
0.720
OSS4
3.364
(1.191)
2.333
0.740
OSS5
3.459
(1.165)
2.241
0.808
Con inuance
In en ion o
Adop
LazzieCha
0.834
0.839
0.627
CUI1
3.256
(1.165)
1.985
0.749
CUI2
3.331
(1.214)
1.859
0.868
CUI3
3.348
(1.227)
1.963
0.753
No e(s): SD: S anda d de ia ion; CR: Composi e eliabili y; AVE: A e age a iance
ex ac ed; VIF: Va iance in la ion ac o
The e alua ion o disc iminan alidi y used Fo nell-La cke c i e ion and The He e o ai -
Mono ai a io o co ela ions (HTMT). Fo nell-La cke c i e ion compa es he co ela ion
be ween he squa e oo o AVE and he la en a iables (Fo nell and La cke , 1981). Table
5 epo s he squa e oo o he AVE exceeded he co ela ions be ween a iables, hus
p o iding suppo o disc iminan alidi y (Hai e al., 2013). The HTMT c i e ion
sugges ed ha he HTMT alues should be ≤ 0.85 (Hensele e al., 2015; F anke and
Sa s ed , 2019). As shown in Table 6, he alues o HTMT we e all lowe han he s ic e
c i e ion o ≤ 0.85 as such we can conclude ha he esponden s unde s ood ha se en
cons uc s a e dis inc (Hensele e al., 2015).
Table 5: Fo nell-La cke c i e ia
ATL
CUI
EMV
EPV
FUV
OSS
SOV
ATL
0.781
CUI
0.553
0.792
EMV
0.603
0.521
0.782
EPV
0.526
0.515
0.549
0.794
FUV
0.534
0.478
0.427
0.371
0.765
OSS
0.552
0.574
0.503
0.452
0.418
0.792
SOV
0.483
0.541
0.372
0.392
0.424
0.449
0.768
No e (s): ATL, A i udes owa d LazzieCha ; CUI, Con inuance in en ion o adop LazzieCha ;
EMV, Emo ional Value; EPV, Epis emic Value; FUV, Func ional Value; OSS, Online shopping Sel -
e icacy; SOV, Social Value
Cha GPT-d i en E-Comme ce AI Cha bo (LazzieCha )
266
Table 6: Disc iminan Validi y (HTMT C i e ion)
ATL
CUI
EMV
EPV
FUV
OSS
SOV
ATL
-
CUI
0.552
-
EMV
0.602
0.521
-
EPV
0.525
0.520
0.550
-
FUV
0.533
0.481
0.430
0.370
-
OSS
0.552
0.571
0.505
0.452
0.420
-
SOV
0.482
0.543
0.371
0.394
0.423
0.447
-
No e (s): ATL, A i udes owa d LazzieCha ; CUI, Con inuance in en ion o adop LazzieCha ; EMV, Emo ional
Value; EPV, Epis emic Value; FUV, Func ional Value; OSS, Online shopping Sel -e icacy; SOV, Social Value
4.5 S uc u al Model
S eukens and Leoi-Weeds (2016) epo ed ha he e was a lack o guidelines o de e mine he
minimum equi ed boo s ap samples, and he e was a signi ican di e ence in he numbe o
boo s ap samples used in PLS-SEM applica ions. Wilcox (2022) sugges s ha a minimum o
2,000 boo s ap samples migh be necessa y o obus analysis. Hence, we employed 5,000
boo s ap esamples o es he hypo heses in his s udy ( e e o Tables 7, 8, 9, and Figu e 3).
Pa h analysis used o in ensi y and di ec ion o he ela ionship be ween a iables is
s udied. The signi icance o he pa h coe icien is es ed o e i y he eliabili y o he
hypo hesis co ela ion. The analysis p ocess ollowed he s a egy p oposed by Hai e al.
(2013). Acco ding o he le el o signi icance, he s a is ical signi icance o he co ela ion
be ween a iables is e alua ed. Table 6 shows he esul s, indica ing ha unc ional alue
(β=0.248, p=0.000), social alue (β=0.184, p=0.009), emo ional alue (β=0.329, p=0.000),
and epis emic alue (β=0.181, p=0.018) all had signi ican e ec s on consume s' a i udes
owa d LazzieCha . Fu he mo e, a i udes owa d LazzieCha (β=0.306, p=0.000)
signi ican ly in luenced consume s' con inuance in en ions o adop LazzieCha . As a
esul , hypo heses H1, H2, H3, H4, and H5 we e all suppo ed.
Table 7: Hypo heses Tes ing Resul s
Hypo hesis
Pa h
Coe icien
ƒ2
- alue
p- alues
Resul s
H1: FUV → ATL
0.248
0.092
3.536
0.000***
Suppo ed
H2: SOV → ATL
0.184
0.052
2.617
0.009**
Suppo ed
H3: EMV → ATL
0.329
0.141
4.130
0.000***
Suppo ed
H4: EPV → ATL
0.181
0.044
2.368
0.018*
Suppo ed
H5: ATA → CUI
0.306
0.114
4.055
0.000***
Suppo ed
No e (s): *p<0.05; **p<0.01; ***p<0.001; ns=nonsigni ican a .05 le el
FUV, Func ional Value; SOV, Social Value; EMV, Emo ional Value; EPV, Epis emic Value;
ATL, A i udes owa d LazzieCha ; CUI, Con inuance in en ion o adop LazzieCha
Yue, Ng & Basha
267
The adjus ed coe icien o de e mina ion alue (R2) ep esen s he p opo ion o a iance
accoun ed o in a gi en model. The esul shows a i udes owa ds LazzieCha (R2 =0.508)
and con inuance in en ion o adop LazzieCha (R2 =0.437) indica ing ha he s uc u al
model ep esen s a good alue o p edic i e accu acy in beha io al esea ch (Hai e al.,
2017). In addi ion o assessing he magni ude o he R² alue as a c i e ion o p edic i e
accu acy, esea che s also equen ly examine S one-Geisse 's Q² alue (S one, 1974;
Geisse , 1974) as a c i e ion o p edic ing co ela ion. By using he blind olding
p ocedu e, he esul s show ha he Q² alues o a i udes owa ds LazzieCha (Q² =0.373)
and con inuance in en ion o adop LazzieCha (Q² =0.344) a e g ea e han ze o which
indica es he s uc u al model's p edic i e ele ance.
Figu e 3. S uc u al Model
4.6 Assessmen o Media ing E ec s
Table 8 show he hypo heses es ing o media ing e ec s esul s which shows consume s'
a i udes owa d LazzieCha as playing a signi ican media ing ole in a ec ing he
Cha GPT-d i en E-Comme ce AI Cha bo (LazzieCha )
268
ela ionship among unc ional alue (β=0.076, p=0.015), social alue (β=0.056 p=0.035),
emo ional alue (β=0.101, p=0.003) and epis emic alue (β=0.055, p=0.046) on
consume s' con inuance in en ion o adop LazzieCha . Thus, all hypo heses o H6 a e
suppo ed based on H6a, H6b, H6c and H6d showed signi ican media ing e ec s. I
con i med ha ull media ion exis s in he da a se (Hai e al., 2013).
Table 8: Hypo heses Tes ing o Media ing E ec s Resul s
Hypo hesis
Pa h
coe icien
- alue
p alues
Resul s
H6a: FUV →ATL →CUI
0.076
2.436
0.015*
Signi ican
H6b: SOV→ATL →CUI
0.056
2.106
0.035*
Signi ican
H6c: EMV→ ATL→CUI
0.101
2.938
0.003**
Signi ican
H6d: EPV →ATL →CUI
0.055
1.993
0.046*
Signi ican
No e (s): *p<0.05; **p<0.01; ***p<0.001; ns=nonsigni ican a .05 le el
FUV, Func ional Value; SOV, Social Value; EMV, Emo ional Value; EPV, Epis emic Value;
ATL, A i ude owa ds LazzieCha ; CUI, Con inuance in en ion o adop LazzieCha
4.7 Assessmen o Mode a ion E ec
Table 9 shows he hypo heses es ing o mode a ing e ec s esul s, in which he indings
indica e ha consume s' online shopping sel -e icacy (β=0.220, alue=3.142, p
alue=0.002<0.05) has a signi ican mode a ing e ec on consume s' a i udes owa d
LazzieCha and consume s' con inuance in en ion o adop LazzieCha . I con i med ha
ull mode a ion exis s in he da a se (Hai e al., 2013).
Table 9: Hypo heses Tes ing o Mode a ing E ec s Resul s
Hypo hesis
Pa h
coe icien
- alue
p- alues
Resul s
H7: OSS x ATL
→CUI
0.220
3.142
0.002**
Signi ican
No e (s): *p<0.05; **p<0.01; ***p<0.001; ns=nonsigni ican a .05 le el
OSS, Online shopping Sel -e icacy; ATL, A i udes owa d LazzieCha ;
CUI, Con inuance in en ion o adop LazzieCha
Fu he mo e, Sma PLS 4 p o ides simple slope plo s in he esul s epo o explain he
mode a ion e ec s eng h ela ionship (Hai e al., 2017). Figu e 4 shows ha o
indi iduals wi h high Online shopping Sel -e icacy (i.e., +1 s anda d de ia ion abo e he
mean; g een line), he e is a s onge ela ionship (i.e., s eepe line) be ween ATL and CUI.
Fo indi iduals wi h low Online shopping Sel -e icacy (i.e., -1 s anda d de ia ion below
he mean; ed line), he slope is la e . Thus, H7 was suppo ed. This means ha compa ed
o consume s wi h low online shopping sel -e icacy, consume s wi h high online shopping
sel -e icacy will mo e s ongly ans o m hei a i udes owa d LazzieCha o hei
con inuance in en ion o adop LazzieCha .
Yue, Ng & Basha
269
Figu e 4. Simple Slope Plo
5. Discussion and Conclusion
Al hough mo e and mo e Cha GPT-d i en cha bo s a e used in he ield o e-comme ce,
li le is known abou he ac o s ha posi i ely a ec consume s' a i ude owa ds his
echnology and hei in en ion o con inue o adop i . P e ious li e a u e shows ha he
ole o consume s' pe cei ed alue in he con ex o e-comme ce cha bo s has no been
ully explo ed, no has he mode a ing ole in he adop ion o he cha bo p ocess (Chen e
al. 2023; Hu e al. 2018). To cla i y his p oblem, based on he p e ious esea ch on e-
comme ce cus ome se ice cha bo s, ou esea ch ex ends he heo y o consume alues,
ollowing he logic o alue-a i ude-beha io , s udies he ac o s ha a ec consume s'
a i ude and in en ion o adop LazzieCha , and conside s he mode a ing ole o online
shopping sel -e icacy in his p ocess. This is s ill a limi ed esea ch ield.
The empi ical esul s con i m he consis ency be ween he indings o his s udy and
p e ious s udies, especially he logical low o pe sonal alues-a i udes-beha io s, which
is an e ec i e s uc u al model o unde s anding and p edic ing beha io in en ions.
A i udes play a key in e media y ole in linking pe sonal alues o beha io , and his
concep has been suppo ed by p e ious li e a u e (Home and Kahle, 1988; Johns on e
al., 2022; Khoi e al., 2018). In o he wo ds, ou esea ch esul s show ha unc ional alue,
social alue, emo ional alue and epis emic alue signi ican ly a ec China’s consume s'
a i udes owa ds LazzieCha . A i udes owa ds LazzieCha play a key media ing ole in
Cha GPT-d i en E-Comme ce AI Cha bo (LazzieCha )
270
he ela ionship be ween consump ion alues (i.e., unc ional alue, social alue, emo ional
alue and epis emic alue) and con inuance in en ion o adop LazzieCha . I is wo h
no ing ha emo ional alue is he s onges p edic o o consume s' a i udes owa ds
LazzieCha , ollowed by unc ional alue, social alue and epis emic alue in adop ing
Cha GPT-d i en e-comme ce AI cha bo s in e-comme ce consume se ices.
Acco ding o he s udy’s indings, he s udy sugges s ha unc ional alue and epis emic
alue se e as impo an d i e s o consume a i udes and in en ions o con inue using
Cha GPT-d i en cha bo s. This inding is consis en wi h p e ious li e a u e in he cha bo
con ex , indica ing ha consume s who pe cei e unc ional alue and possess a cu iosi y
o no el y and new knowledge a e mo e likely o o m posi i e a i udes and adop cha bo
se ices (Cicco e al., 2020; Za ouali e al., 2018). Resul s con i m ha Cha GPT-d i en e-
comme ce cha bo s can o e highe esponsi eness wi h high accu acy, educing consume
e o and ime while p o iding knowledge in o ma ion abou p oduc s. This e iciency-
d i en unc ion and consume s' cu iosi y hen a ec he o ma ion o a i udes and
in en ions. This s udy ound ha social alue demons a es a signi ican impac on
consume s' a i udes owa d LazzieCha , subsequen ly a ec ing hei in en ions o con inue
using i . This esul is simila o p e ious esea ch ha highligh s he in luence o social
alue on use a i udes and in en ions owa d adop ing digi al echnology se ices
(Chaouali e al., 2023; Kau e al., 2021; Khoi e al., 2018; Omigie e al., 2017; Wu e al.,
2018). Fo example, Khoi e al. (2018) epo ed ha social alues in luence Vie namese
consume s' a i udes and hei in en ion o adop mobile comme ce se ices while calling
o paying a en ion o explo ing de eloping coun y consume alues o unde s and he
adop ion o mobile comme ce echnology. Hence, his s udy is a supplemen o his, which
u he con i ms he p e ious esea ch esul s o con i m he impo ance o consume s'
social alue o hei a i ude. In addi ion, he esul s suppo ha consume s' a i ude
owa ds LazzieCha has a signi ican impac on hei con inuance in en ion o adop
LazzieCha . I is simila o pas li e a u e ha shows ha a i udes posi i ely di ec ly
in luence beha io in en ion (Johns on e al. 2022; Kuma e al. 2024). Fo example, Kuma
e al. (2024) emphasized ha use 's a i udes di ec ly in luence hei in en ions o adop
Cha GPT.
Acco ding o he s udy’s indings, online shopping sel -e icacy con i ms i s signi ican
mode a ing ole be ween a i udes owa d LazzieCha and con inuance in en ions o adop
i . This inding was simila o pas li e a u e (Liu e al., 2017; Yi and Gong, 2008; Zha e
al., 2013). Fo example, Zha e al. (2013) ound ha online shopping sel -e icacy s ongly
mode a es he e ec s o consume s' online shopping beha io and decision p ocess. This
means ha compa ed wi h consume s wi h low online shopping sel -e icacy, consume s
wi h high online shopping sel -e icacy a e mo e capable o e ec i ely ans o ming hei
a i udes o hei con inuance in en ion o adop Cha GPT-d i en e-comme ce cha bo s in
e-comme ce se ices.
Yue, Ng & Basha
271
5.1 Theo e ical Implica ions
Theo e ically, he in es iga ion o his s udy can be used as new in o ma ion o suppo
Cha GPT in business knowledge. The concep o Cha GPT is s ill in he e olu iona y s age
(Kuma e al., 2024). The pu pose o his s udy is o explo e he mo i a ion ac o s ha
a ec consume s' con inuous in en ion o adop Cha GPT-d i en cha bo s in he e-
comme ce consume se ice con ex . I ills a majo esea ch gap by s udying consume s'
con inuous in en ion o adop Cha GPT-d i en e-comme ce cha bo s om he pe spec i e
o consume alues, which is a pe spec i e ha is o en o e looked in exis ing esea ch
(Hu e al. 2018). P e ious s udies o e-comme ce cha bo s mainly ocused on
pe soni ica ion (Sheehan e al. 2020), con inued use (Li e al. 2021), emo ional a achmen
(Lee e al. 2021), us (Cheng e al. 2021), pe cei ed decision quali y (Chen e al. 2020)
and he ela ionship be ween cus ome loyal y and se ice quali y (Chen e al. 2023). The e
is limi ed explo e he ole o consump ion alues in he con ex o AI e-comme ce cha bo s.
In he Cha GPT con ex , mos esea che s ha e explo ed Cha GPT om educa ional and
echnological pe spec i es, o example, applica ions and backg ounds (Pasupule i and
Thiyyagu a, 2023) and managemen concep s and heo ies (Ko zynski e al. 2023).
Besides, he pas li e a u e s udies on Cha GPT used in he e-comme ce con ex ha e been
commen a y-based s udies, such as (Kshe i, 2024), o e looking explo a o y s udies ha
could p o ide deepe insigh s in o po en ial ac o s in luencing he adop ion o Cha GPT-
d i en e-comme ce cha bo . Cha GPT-d i en e-comme ce cha bo s ha e g ea po en ial,
he e a e huge heo e ical implica ions o unde s and how alue-based consume -d i en
mo i es a ec hei con inuance in en ions o adop Cha GPT-d i en e-comme ce cha bo s
in e-comme ce consume se ices. Hence, his pape can be used as he basic amewo k
o u u e esea ch o gain u he insigh s and in o ma ion on Cha GPT-d i en e-
comme ce cha bo s.
In addi ion, his s udy p o ed he applicabili y o consump ion alues wi hin he alue-
a i ude-beha io model. Ou esea ch esul s con i m ha emo ional alue and cogni i e
alue a e he main d i e s o consume s' a i udes owa ds LazzieCha , ollowed by
unc ional alue, social alue and epis emic alue, which indi ec ly a ec con inuance
in en ion o adop LazzieCha h ough hei a i udes. This s udy is help ul o he heo e ical
p og ess and applicabili y con ex o he li e a u e by in eg a ing he heo y o consump ion
alue wi h his model, which can be ex ended o explo e and p edic he in luence o
a ious alues on consume a i udes and in en ions owa d adop ing any a i icial
in elligence cha bo se ice. A he same ime, his s udy emphasizes he mode a ing e ec
o consume s' online shopping sel -e icacy be ween consume s 'a i udes and hei
con inuance in en ions o adop Cha GPT-d i en cha bo s in e-comme ce se ices. This
expands he exis ing li e a u e on he egula o y e ec s o consume cha ac e is ics, such
as online shopping sel -e icacy in he con ex o using Cha GPT-d i en cha bo s in
eme ging echnologies and e-comme ce.
Cha GPT-d i en E-Comme ce AI Cha bo (LazzieCha )
272
5.2 Manage ial Implica ions
The indings o his esea ch o e se e al implica ions o he in eg a ion o Cha GPT-
d i en cha bo s in o he e-comme ce indus y. Fi s ly, hese s udy esul s p o ide aluable
insigh s in o consume s' pe cei ed consump ion alue when adop ing Cha GPT-d i en
cha bo s in e-comme ce se ices. This unde s anding can guide ac ionable s a egies o
Cha GPT's applica ion in he speci ic domain o e-comme ce consume se ice. Fo
ins ance, he high alue placed on emo ional, unc ional, social, and epis emic alues
indica es ha he inno a i e echnologies and ea u es o e ed by Cha GPT-d i en cha bo s
hold s ong appeal o consume s. This can con ibu e o enhancing consume s' emo ional
sa is ac ion, imp o ing hei social connec ions and image, and s imula ing hei cu iosi y.
Manage s ope a ing in he e-comme ce and consume se ice sec o s can le e age
Cha GPT-d i en cha bo s o enhance hei e-comme ce consume suppo . They should
ocus on ope a ional excellence and explo e pionee ing consume se ice app oaches o
deli e emo ional, unc ional, social, and epis emic alues, ul ima ely enhancing he
o e all e-comme ce consume se ice expe ience.
Secondly, ou indings emphasize he signi icance o a i ude as he p ima y di ec d i e
o consume s' con inuance in en ions o adop LazzieCha . This shows ha consume s'
a i udes play a key ole in hem adop new digi al echnology se ices. As he use o
Cha GPT-d i en cha bo s in e-comme ce se ices ep esen s a no el and inno a i e
me hod, compa ed wi h adi ional cha bo s, manage s should gi e p io i y o se ice
quali y and p o ide aluable in o ma ion, o c ea e a unique consume se ice expe ience.
This me hod can posi i ely in luence people's a i udes, and hus lead consume s o be mo e
willing o adop his new echnology o mee hei e-comme ce consume se ice demand.
Thi dly, he esea ch esul s emphasize he mode a ing e ec o online shopping sel -
e icacy be ween consume 'a i udes and hei con inuance in en ion o adop LazzieCha .
This means ha consume s wi h high sel -e icacy in online shopping a e mo e likely o
use Cha GPT-d i en cha bo s in e-comme ce se ices. E-comme ce consume se ice
p o ide s and manage s should ake measu es o enhance consume s' powe by p o iding
clea calls o ac ion and guidance ips on online consume se ice pla o ms. This can be
achie ed h ough a use - iendly websi e design, in ui i e na iga ion and quick cus ome
suppo . In addi ion, online e aile s can guide consume s on how o in e ac e ec i ely
wi h Cha GPT-d i en cha bo s by p o iding u o ials o equen ly asked ques ions. Such
measu es will p omo e he adop ion o Cha GPT-d i en cha bo s in e-comme ce se ices
and imp o e he o e all consume expe ience.
Finally, i is possible o imp o e he e iciency and quali y o e-comme ce consume
se ice by using cha bo s d i en by Cha GPT in e-comme ce se ice. Howe e , Cha GPT
is biased agains ce ain cul u al and language combina ions, which leads o inapp op ia e
o biased esponses, and i may ake addi ional ine- uning and expe imen s o achie e he
co ec balance be ween con ol and quali y (Kuma e al., 2024). Hence, unde s anding
consume s' consump ion alue ac o s ha in luence hei con inuance in en ion o adop
Yue, Ng & Basha
279
Lee, C.T., Pan, L.Y. and Hsieh, S.H. (2021). A i icial in elligen cha bo s as b and
p omo e s: a wo-s age s uc u al equa ion modeling-a i icial neu al ne wo k
app oach. In e ne Resea ch, 32(4), 1329-1356.
Lei, S. I., Shen, H., & Ye, S. (2021). A compa ison be ween cha bo and human se ice:
Cus ome pe cep ion and euse in en ion. In e na ional Jou nal o Con empo a y
Hospi ali y Managemen , 33(11), 3977–3995.
Li, L., Lee, K. Y., Emokpae, E., & Yang, S.-B. (2021). Wha makes you con inuously use
cha bo se ices? E idence om chinese online a el agencies. Elec onic Ma ke s, 31(3),
575–599.
Li, Y., Xu, Z., & Xu, F. (2018). Pe cei ed con ol and pu chase in en ion in online
shopping: The media ing ole o sel -e icacy. Social Beha io and Pe sonali y: an
in e na ional jou nal, 46(1), 99-105.
Liu, J., Cho, S., & Pu a, E. D. (2017). The mode a ing e ec o sel -e icacy and gende
on wo k engagemen o es au an employees in he Uni ed S a es. In e na ional Jou nal
o Con empo a y Hospi ali y Managemen , 29(1), 624–642.
Lund, B. D., & Wang, T. (2023). Cha ing abou Cha GPT: How may AI and GPT impac
academia and lib a ies? Lib a y Hi Tech News, 40(3), 26–29.
Kang, J.,Y., M. (2014). Augmen ed eali y and mo ion cap u e appa el e-shopping alues
and usage in en ion. In e na ional Jou nal o Clo hing Science and Technology, 26(6),
486-499.
Min, S., So, K.K.F. and Jeong, M. (2019). Consume adop ion o he Ube mobile
applica ion: insigh s om di usion o inno a ion heo y and echnology accep ance
model. Jou nal o T a el and Tou ism Ma ke ing, 36(7), 770-783.
Naba i, A., Tagha i-Fa d, M.T., Hana izadeh, P. and Tagh a, M.R. (2016). In o ma ion
echnology con inuance in en ion: a sys ema ic li e a u e e iew. In e na ional Jou nal o
E-Business Resea ch, 12(1), 58-95.
Nau iyal, R., Alb ech , J. N., & Nau iyal, A. (2023). Cha GPT and ou ism academia.
Annals o Tou ism Resea ch, 99, 103544.
Ne emeye , R. G., Bea den, W. O., & Sha ma, S. (2003). Scaling P ocedu es: Issues and
Applica ions. Thousand Oaks, CA: Sage Publica ions.
Nunnally, J.C. and Be ns ein, I.H. (1994) The Assessmen o Reliabili y. Psychome ic
Theo y, 3, 248-292.
Omigie, N.O., Zo, H., Rho, J.J. and Ciganek, A.P. (2017). Cus ome p e-adop ion choice
beha io o M-PESA mobile inancial se ices: Ex ending he heo y o consump ion
alues. Indus ial Managemen & Da a Sys ems, 117(5), 910-926.
Cha GPT-d i en E-Comme ce AI Cha bo (LazzieCha )
280
Ou lookindia. (2023.). Ai India Makes $200 Million Ini ial In es men o Digi al Sys ems
Mode nisa ion; To Use Cha GPT-D i en Cha bo . Re ie ed 13 Augus 2023, om
h ps://www.ou lookindia.com/business/ai -india-makes-200-million-ini ial-in es men -
o -digi al-sys ems-mode nisa ion- o-use-cha gp -d i en-cha bo -news-280919
Panda, S., & Kau , N. (2023). Explo ing he iabili y o Cha GPT as an al e na i e o
adi ional cha bo sys ems in lib a y and in o ma ion cen e s. Lib a y Hi Tech News, 40(3),
22–25.
Pasupule i, R.S., Thiyyagu a, D. (2024). An empi ical e idence on he con inuance and
ecommenda ion in en ion o Cha GPT among highe educa ion s uden s in India: An
ex ended echnology con inuance heo y. Educa ion and In o ma ion Technologies .
h ps://doi.o g/10.1007/s10639-024-12573-7
Peng, K.F., Chen, Y. and Wen, K.W. (2014). B and ela ionship, consump ion alues and
b anded app adop ion. Indus ial Managemen and Da a Sys ems, 114(8), 1131-1143.
Rahi, S., O hman Mansou , M.M., Alha a sheh, M. and Alghizzawi, M. (2021). The pos -
adop ion beha io o in e ne banking use s h ough he eyes o sel -de e mina ion heo y
and expec a ion con i ma ion model. Jou nal o En e p ise In o ma ion Managemen , 34
(6), 1874-1892.
Roge s, E.M. (1995), Di usion o Inno a ion, 4 h Ed., F ee P ess, New Yo k, NY.
Roig, J.C.F., Ga cia, J.S., Tena, M.A.M, & Llo ens Monzonis, J. (2006). Cus ome
pe cei ed alue in banking se ices. In e na ional Jou nal o Bank Ma ke ing, 24(5), 266–
283.
Se iceBell. (2023). 53 Cha bo S a is ics Fo 2022: Usage, Demog aphics, T ends.
Se iceBell. Online a ailable a : h ps://www.se icebell.com/pos /cha bo -
s a is ics#:~: ex =68%25%20o %20use s%20enjoy%20 he,%24100%20billion%20in%2
0ecomme ce%20 ansac ions.
Shawa BA, A well E. (2007) Di e en measu emen s me ics o e alua e a cha bo
sys em. In: P oceedings o he Wo kshop on B idging he Gap: Academic and Indus ial
Resea ch in Dialog Technologies; 89-96.
Sheehan, B., Jin, H.S. and Go lieb, U. (2020). Cus ome se ice cha bo s:
an h opomo phism and adop ion. Jou nal o Business Resea ch, 115, 14-24.
She h, J.N., Newman, B.I. and G oss, B.L. (1991). Why we buy wha we buy: a heo y o
consump ion alues. Jou nal o Business Resea ch, 22(2), 159‐70.
Sidaoui, K., Jaakkola, M., & Bu on, J. (2020). AI eel you: cus ome expe ience
assessmen ia cha bo in e iews. Jou nal o Se ice Managemen , 31(4), 745-766.
Simila Web. (2023). T ack he websi e pe o mance o lazada.sg. Re ie ed 15 Augus
2023, om h ps://www.simila web.com/zh/websi e/lazada.sg/#o e iew
Yue, Ng & Basha
281
S one, M. (1974). C oss- alida ion and mul inomial p edic ion. Biome ika, 61(3), 509-
515.
S ai s Times. (2021). 1 in 2 companies in Singapo e has sped up AI oll-ou in he wake
o Co id-19: S udy. The S ai s Times. h ps://www.s ai s imes.com/ ech/ ech-news/1-in-
2-companies-in-singapo e-has-sped-up-ai- oll-ou -in- he-wake-o -co id-19-s udy
S eukens, S., & Le oi-We elds, S. (2016). Boo s apping and PLS-SEM: A s ep-by-s ep
guide o ge mo e ou o you boo s ap esul s. Eu opean Managemen Jou nal, 34(6),
618–632.
S zelecki, A. (2023). To use o no o use Cha GPT in highe educa ion? A s udy o
s uden s’ accep ance and use o echnology. In e ac i e Lea ning En i onmen s, 1-14,
doi: 10.1080/10494820.2023.2209881.
Suki, M. N. (2013). Young consume ecological beha iou : he e ec s o en i onmen al
knowledge, heal hy ood, and heal hy way o li e wi h he mode a ion o gende and
age. Managemen o En i onmen al Quali y: An In e na ional Jou nal, 24(6), 726-737.
Sweeney, J.C. and Sou a , G.N. (2001). Consume pe cei ed alue: he de elopmen o a
mul iple i em scale. Jou nal o Re ailing, 77(2), 203-220.
Tandon, U. (2023). Cha bo s, i ual- y-on (VTO), e-WOM: Modeling he de e minan s
o a i ude’ and con inued in en ion wi h PEEIM as mode a o in online shopping. Global
Knowledge, Memo y and Communica ion. [ahead o p in ].
T i edi, J. (2019). Examining he cus ome expe ience o using banking Cha bo s and i s
impac on b and lo e: he mode a ing ole o pe cei ed isk. Jou nal o In e ne Comme ce,
18(1), 91-111.
Tu el, O., Se enko, A. and Bon is, N. (2010). Use accep ance o hedonic digi al a e ac s:
a heo y o consump ion alues pe spec i e. In o ma ion and Managemen , 47(1), 53-59.
Wang, W., Cao, D., & Ameen, N. (2023). Unde s anding cus ome sa is ac ion o
augmen ed eali y in e ail: A human alue o ien a ion and consump ion alue pe spec i e.
In o ma ion Technology & People, 36(6), 2211–2233.
Wilcox, R. R. (2022). In e ences in he One-Sample Case. In oduc ion o Robus
Es ima ion and Hypo hesis Tes ing (pp.107–151). Else ie .
Wu, S.-I. and Chang, H.-L. (2016). The model o ela ionship be ween he pe cei ed
alues and he pu chase beha io s owa d inno a i e p oduc s. Jou nal o Managemen
and S a egy, 7(2), 31-45.
Wu, W., Huang, V., Chen, X., Da ison, R.M. and Hua, Z. (2018). Social alue and online
social shopping in en ion: he mode a ing ole o expe ience. In o ma ion Technology and
People, 31(3), 688-711.
Cha GPT-d i en E-Comme ce AI Cha bo (LazzieCha )
282
Xu, Y., Jeong, E., Baiomy, A.E. and Shao, X. (2020). In es iga ing onsi e es au an
in e ac i e sel -se ice echnology (ORISST) use: cus ome expec a ions and
in en ions. In e na ional Jou nal o Con empo a y Hospi ali y Managemen , 32(10), 3335-
3360.
Yan, M., Filie i, R., & Go on, M. (2021). Con inuance in en ion o online echnologies: A
sys ema ic li e a u e e iew. In e na ional Jou nal o In o ma ion Managemen , 58,
102315.
Yang, H., & Li, D. (2021). Unde s anding he da k side o gami ica ion heal h
managemen : A s ess pe spec i e. In o ma ion P ocessing & Managemen , 58(5),
102649.
Yang, H.-L., & Lin, R.-X. (2017). De e minan s o he in en ion o con inue use o SoLoMo
se ices: Consump ion alues and he mode a ing e ec s o o e loads. Compu e s in
Human Beha io , 73, 583–595.
Yang, Y., Liu, Y., Li, H. and Yu, B. (2015). Unde s anding pe cei ed isks in mobile
paymen accep ance. Indus ial Managemen and Da a Sys ems, 115(2), 253-269.
Yi, Y. and Gong, T. (2008). The elec onic se ice quali y model: he mode a ing e ec o
cus ome sel -e icacy. Psychology & Ma ke ing, 25(7), 587-601.
Yoon, J. and Yu, H. (2022). Impac o cus ome expe ience on a i ude and u iliza ion
in en ion o a es au an -menu cu a ion cha bo se ice. Jou nal o Hospi ali y and Tou ism
Technology, 13(3), 527-541.
Yo dan, J. (2023). Lazada launches Cha GPT-powe ed cha bo . Techinasia. A ailable a :
h ps://www. echinasia.com/lazada-launches-ecomme ce-ai-cha bo -powe ed-cha gp
Za ouali, B., B oeck, E.V., Wal a e, M. and Poels, K. (2018). P edic ing consume
esponses o a cha bo on acebook. Cybe psychology, Beha io , and Social Ne wo king,
21(8), 491-497.
Zha, X., Li, J., & Yan, Y. (2013). In o ma ion sel -e icacy and in o ma ion channels:
Decision quali y and online shopping sa is ac ion. Online In o ma ion Re iew, 37(6), 872–
890.
Zikmund, W. G., Ca , J. C., & G i in, M. (2012). Business esea ch me hods. Aus alia:
CengageB ain.com.
Zums ein, D., & Hunde ma k, S. (2018). Cha bo s: an in e ac i e echnology o
pe sonalized communica ion and ansac ion. IADIS In e na ional Jou nal on
WWW/In e ne , 15(1), 96-109.
Yue, Ng & Basha
283
Annexu e
Table 1: Measu emen I ems
Func ional Value (FUV) (Roig e al., 2006)
FUV1
I hink ha LazzieCha knows i s job well.
FUV2
I hink ha LazzieCha can o e knowledge in o ma ion ha is up o da e.
FUV3
I hink ha he in o ma ion p o ided by LazzieCha has always been e y
aluable o me.
FUV4
I hink ha LazzieCha o e high-quali y se ice.
Social Value (SOV) (Omigie e al.,2017)
SOV1
I hink ha using LazzieCha will show my be e social image o o he s.
SOV2
I hink ha using LazzieCha can p o ide eliable se ices and hones
opinions.
SOV3
I hink ha using LazzieCha inc eases my social ela ionship wi h amily,
iends, g oups, associa ions, and so on.
Emo ional Value (EMV) (Ka jaluo o e al., 2021)
EMV1
I can use LazzieCha a any ime.
EMV2
I hink ha LazzieCha enables me o exp ess my pe sonali y.
EMV3
I hink ha LazzieCha makes me eel ashionable.
EMV4
I hink ha LazzieCha helps me li e and wo k sa is ac o ily.
Epis emic Value (EPV) (Omigie e al.,2017)
EPV1
I s a ed using LazzieCha because I was cu ious abou i s se ice.
EPV2
I s a ed using LazzieCha because o discussions and ecommenda ions
om people a ound me.
EPV3
I s a ed using LazzieCha because I wan ed o lea n a new li es yle.
A i udes owa d LazzieCha (ATL) (Han e al., 2017)
ATL1
I am in e es ed in using LazzieCha .
ATL2
I would like o ea ecommenda ions by LazzieCha as one o my online
pu chase decisions.
ATL3
I gi es me a posi i e eeling o use LazzieCha .
Online shopping Sel -e icacy (OSS) (Dash & Saji, 2008)
OSS1
I eel con iden communica ing wi h LazzieCha i he e a e clea
ins uc ions o my e e ence.
OSS2
I eel con iden communica ing wi h LazzieCha e en i no one ells me
how o use i .
OSS3
I eel con iden communica ing wi h LazzieCha e en hough I ha e ne e
expe ienced he same be o e.
OSS4
I eel con iden communica ing wi h LazzieCha e en i I ha e jus seen
someone using i be o e ying i mysel .
OSS5
I eel con iden communica ing wi h LazzieCha i I ha e jus he online
help unc ion o assis ance.
Cha GPT-d i en E-Comme ce AI Cha bo (LazzieCha )
284
Con inuance in en ion o adop LazzieCha (CUI) (Choudhu y and
Shamsza e,2023)
CUI1
I would use LazzieCha o online shopping- ela ed que ies.
CUI2
I would ake decisions based on he ecommenda ions p o ided by
LazzieCha .
CUI3
I would con inue using LazzieCha in u u e.
Table 2: Demog aphic P o ile o 305 Su ey Responden s
I ems
F equency
(N=305)
Pe cen
(%)
Gende
Male
190
62.30
Female
115
37.70
Age
18-25
52
17.00
26-35
163
53.40
36-45
67
22.00
>45
23
7.50
Educa ion le el
Doc o a e
17
5.60
Mas e ’s deg ee
83
26.20
Bachelo ’s deg ee
125
41.00
O he
80
26.20
Indi idual mon hly income (SGD$)
Unde 2,500
44
14.40
2,501-5,000
124
40.70
5,001-7,500
110
36.10
7501 o highe
27
8.90
Hou s spend on LazzieCha (pe week)
Less han 3 hou s
66
21.60
3-6 hou s
122
40.00
7-9 hou s
93
30.50
Mo e han 9 hou s
24
7.90
