Bolos, Ma cel Ioan e al.
A icle
AI cha bo s: Fas acking sus ainabili y epo analysis o
enhanced decision making
Am i ea u Economic
P o ided in Coope a ion wi h:
The Bucha es Uni e si y o Economic S udies
Sugges ed Ci a ion: Bolos, Ma cel Ioan e al. (2024) : AI cha bo s: Fas acking sus ainabili y epo
analysis o enhanced decision making, Am i ea u Economic, ISSN 2247-9104, The Bucha es
Uni e si y o Economic S udies, Bucha es , Vol. 26, Iss. Special Issue No. 18, pp. 1241-1255,
h ps://doi.o g/10.24818/EA/2024/S18/1241
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/319802
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/
New T ends in Sus ainable Business and Consump ion
AE
Vol. 26 • Special Issue No 18 • No embe 2024 1241
AI CHATBOTS: FAST TRACKING SUSTAINABILITY REPORT ANALYSIS
FOR ENHANCED DECISION MAKING
Ma cel Ioan Boloș1, Ș e an Rusu2, Claudia Diana Sabău-Popa3,
Dana Simona Ghe ai4, Ad ian Neg ea5 and Mihai-Ioan C ișan6
1) Uni e si y o O adea, Facul y o Economic Sciences, O adea, Romania
2) Uni e si y o O adea, Doc o al School o Economic Sciences, O adea, Romania
3,4,5)Uni e si y o O adea, Facul y o Economic Sciences, O adea, Romania
6)Babeș-Bolyai Uni e si y, Facul y o Economics and Business Adminis a ion,
Cluj-Napoca, Romania
Please ci e his a icle as:
Boloș M. I., Rusu Ș., Sabău-Popa C. D., Ghe ai D. S.,
Neg ea A. and C ișan M.-I. AI Cha bo s: Fas T acking
Sus ainabili y Repo Analysis o Enhanced Decision
Making. Am i ea u Economic, 26(Special Issue No. 18),
pp. 1241-1255.
DOI: h ps://doi.o g/10.24818/EA/2024/S18/1241
A icle His o y
Recei ed: 8 Augus 2024
Re ised: 11 Sep embe 2024
Accep ed: 9 Oc obe 2024
Abs ac
This pape explo es he in eg a ion o a i icial in elligence (AI) in businesses, ocusing on
he u ili y o cha bo s o sus ainabili y epo analysis. Using Cha GPT echnology ia he
Cha -based pla o m, we de eloped a pe sonalised cha bo o ex ac da a om sus ainabili y
epo s, wi h he aim o acili a ing decision-making p ocesses. The s udy includes a li e a u e
e iew on AI's impac on he economy and labou ma ke s, ollowed by a me hodology base
on echnology Cha GPT de ailing cha bo de elopmen and es ing using a sus ainabili y
epo o a company lis ed on he Romanian s ock ma ke . The esul s demons a e he
e icacy in p o iding accu a e inancial insigh s, o e ing po en ial bene i s o analys s,
in es o s, and business o ganisa ions manage s. By ha nessing AI-powe ed cha bo s,
o ganisa ions can s eamline ope a ions and gain a compe i i e edge in oday's digi al
landscape.
Keywo ds: a i icial in elligence; cha bo ; inance; inancial analysis; decision-making
JEL Classi ica ion: C69, C89, G19, G29, G39
Au o de con ac , Claudia Diana Sabău-Popa - email: dianasabaupo[email p o ec ed]
Aces a es e un a icol cu acces deschis dis ibui în con o mi a e cu e menii C ea i e
Commons A ibu ion License (h ps://c ea i ecommons.o g/licenses/by/4.0), ca e pe mi e
u iliza ea, dis ibui ea și ep oduce ea ă ă es icții în o ice mediu, cu condiția ca luc a ea
o iginală să ie ci a ă co ec . © 2023 Toa e d ep u ile apa țin au o ilo .
AE
AI Cha bo s: Fas T acking Sus ainabili y Repo Analysis
o Enhanced Decision Making
1242 Am i ea u Economic
In oduc ion
The de elopmen o ad anced echnologies in he ope a ionalisa ion o business
o ganisa ions' ac i i ies, wi h he help o a i icial in elligence (AI), can lead o an inc ease
in hei pe o mance by inc easing employee p oduc i i y, op imising sales, sa ing ene gy
use, e c. In a digi ised e a, i is impe a i e ha all o ganisa ions employees ha e digi al skills,
so o ganisa ions, ega dless o hei size, should in es in he de elopmen and con inuous
aining o hei employees' skills as a p e equisi e o ensu ing u u e pe o mance.
AI has a majo in luence on decision-making pa e ns in companies and on ede ining he
asks o hei employees, equi ing con inuous adap a ion o he ad ancemen o a i icial
in elligence echniques (Thomas e al., 2016). The same idea is also suppo ed by Ag awal
e al. (2019) who show in hei esea ch he use ulness o hinking in e ms o p edic ion asks
and decision asks, p edic ion ha ing no alue wi hou decision, AI eplacing employees in
he wo kplace depending on he deg ee o which he basic skills in ol ed in he in ended job
in ol e p edic ion.
We a e seeing an inc easing amoun o En i onmen al, Social, and Go e nance ac o s (ESG)
in luence on in es men decisions a he in es o le el. ESG-awa e in es ing bene i s
in es o s bo h inancially and non- inancially and p omo es social and en i onmen al
esponsibili y. e al. Sul ana (2018). The s udy's indings encou age businesses o adop
en i onmen ally iendly p ac ices, sound social and go e nance policies in o de o suppo
a mo e sus ainable economy, and au ho i ies may u ilise his da a o c ea e ESG- ela ed
legisla ion ha p ese es social and en i onmen al equilib ium in he s ock ma ke .
Ad anced a i icial in elligence/machine lea ning ools ha a e inc easingly applied in he
inancial sec o a e able o pe o m clea asks ha p e iously equi ed human in elligence.
Fo his eason, inancial ins i u ions ely mo e and mo e on a i icial in elligence/machine
lea ning ools o asse managemen , op imisa ion o cus ome expe ience, algo i hmic
ading, op imal isk managemen , op imisa ion o lending ope a ions (Belhaj, & Hachaïchi,
2023).
In his ega d, he In e na ional Mone a y Fund (2021) in he s udy on he use o a i icial
in elligence in he inancial sec o poin s ou ha sys ems based on a i icial
in elligence/machine lea ning (AI/ML) ha e made signi ican p og ess in ecen yea s,
AI/ML sys ems a e al eady able o pe o m asks well-de ined ha usually equi e human
in elligence. I is es ima ed ha he accele a ed de elopmen and p opelled by he pandemic
c isis o he digi isa ion o employee asks h ough he applica ion o a i icial in elligence
will gene a e massi e ans o ma ions in he labou ma ke , om job equi emen s o ask
design, as well as employee wo k e alua ion (C ama enco e al., 2023).
In he con ex o digi al ans o ma ions, o ganisa ions ha e adop ed echnological
inno a ions based on a i icial in elligence and ecen s udies ha e shown he abili y o
a i icial in elligence applied in o ganiza ions o imp o e hei pe o mance bo h a he
o ganisa ional le el ( inancial, ma ke ing and adminis a i e) and a he p ocess le el, as well
as inc easing he e u n on in es men in AI- ans o med p ojec s (CIGREF, 2018; C ews,
2019; Wamba-Taguimdjee al., 2020)
New T ends in Sus ainable Business and Consump ion
AE
Vol. 26 • Special Issue No 18 • No embe 2024 1243
In his pape , we aimed o show he use ulness o cha bo s in he cu en ac i i y o
o ganisa ions, by c ea ing a cha bo based on he echnology behind Cha GPT, using he
Cha base pla o m in o de o ob ain selec i e inancial in o ma ion o business analysis.
The cha bo ha we de elop and p opose in his pape can signi ican ly educe he ime
equi ed o sea ch o speci ic in o ma ion and hus speed up he analysis o inancial epo s
and imp o e he decision-making p ocess.
This a icle is s uc u ed in h ee sec ions: he i s sec ion analyses he specialized li e a u e
ega ding bo h cha bo s and hei u ili y in he economy and he ans o ma ions in he labou
ma ke h ough he in eg a ion o a i icial in elligence in he ac i i y o companies; he
second sec ion p esen s he me hodology o building a cha bo based on Cha GPT
echnology, while he las sec ion highligh s he esul s ob ained and he implica ions o
in es o s, inancial analys s, and company manage s.
1. Li e a u e e iew
C ama enco e al. (2024) sys emic li e a u e e iew encompassing 639 pape s highligh s he
challenges posed by AI-d i en dis up ion. The indings unde sco e he need o aining,
mainly o enhance and eskilling ini ia i es o add ess skill equi emen s in a wo ld in which
AI is used. The need o p ope egula ion and p oac i e e hical and esponsible in eg a ion
o AI is emphasised, as well as o ensu e holis ic employee well-being and sound p o essional
de elopmen in an e e -changing wo ld.
The esea ch pape published by E ns e al. (2018) examines he impac o AI on labou
ma ke s, compa ing i wi h p e ious au oma ion wa es. I iden i ies oppo uni ies o
p oduc i i y gains, especially in de eloping coun ies, bu also highligh s he isks o
exace ba ing inequali y. The conclusions emphasise he signi ican po en ial o AI
echnologies ac oss sec o s and skill le els, bu highligh he need o encompassing policies
o ensu e equi able dis ibu ion o bene i s. The collabo a ion be ween policy make s and
s akeholde o add essing ma ke concen a ion, p o ec ing da a igh s, and os e ing
in e na ional coope a ion, along wi h cons an moni o ing and egula ions o add ess e hical
issues and ensu ing socie al-cen ic AI a e some o he ecommenda ions o e ed by he
au ho s. Simila ly, Beljah e al. (2023) emphasises he necessi y o policymake s o balance
he isk and bene i s o AI adop ion using obus egula o y esponses, calling o enhanced
o e sigh o mi iga e po en ial eme ging isks and ensu e e hical and esponsible AI
inno a ion, hus p o ec ing inancial s abili y and consume wel a e.
ESG epo ing is an impo an conside a ion when making in es men decisions. In hei
esea ch, Sul ana e al. (2018) in es iga e indi idual s ock ma ke in es o s' p e e ences o
En i onmen al, Social, and Go e nance (ESG) issues and how hese p e e ences, in addi ion
o in es ing in en , in luence decision-making. The s udy also examines he mode a ing ole
o he in es men ho izon, speci ically how he long- e m iew a ec s he ela ionship
be ween ESG issues and in es men decisions. The s udy's key indings demons a e ha
social, en i onmen al, and go e nance conce ns ha e a signi ican impac on in es o
decision-making p ocesses. Long- e m in es o s a e mo e willing o examine ESG p oblems
as a means o educing isk and ensu ing a sus ainable e u n.
AE
AI Cha bo s: Fas T acking Sus ainabili y Repo Analysis
o Enhanced Decision Making
1244 Am i ea u Economic
New T ends in Sus ainable Business and Consump ion
AE
Vol. 26 • Special Issue No 18 • No embe 2024 1245
Fu he mo e, in hei s udy, Pa k and Jan (2021) emphasise ha ESG has a majo impac on
in es o decisions, pa icula ly on en i onmen al and go e nance issues. The s udy looks a
how ins i u ional in es o s inco po a e ESG ac o s in o hei own in es ing decisions, bo h
globally and speci ically in Sou h Ko ea. They concluded in hei analysis ha ins i u ional
in es o s wi h o wa d- hinking in es men s, such as pension unds, place a highe weigh
on go e nance issues such as owne ship igh s and CEO epu a ion. On he con a y, sho -
e m in es o s p io i ise consume happiness and en i onmen ally iendly s a egies.
Companies ha c ea e a coun y-speci ic ESG amewo k can imp o e compliance, enhance
ma ke epu a ion, and a ac mo e in es men , indica ing sus ainabili y and long- e m
g ow h po en ial o bo h domes ic and in e na ional in es o s. Fu he mo e, he s udy
highligh s ha including coun y-speci ic c i e ia in a b oade global ESG amewo k can
lead o mo e accu a e p edic ions o business pe o mance.
These heo ies a e ein o ced and co obo a ed by Mehwish e al. (2022), who conduc ed a
ecen s udy on how ESG a ec s indi idual in es men decisions a he Pakis an S ock
Exchange. They applied a model based on he Theo y o Planned Beha iou . This s udy
disco e ed ha PSX in es o s place a highe alue on go e nance issues such as sha eholde
igh s and co po a e e hics han on social o en i onmen al conce ns. This p e e ence
highligh s he ole o excellen go e nance in ec ui ing and main aining in es o s. The s udy
also disco e ed ha e ec i e ESG in eg a ion can boos a company's epu a ion and inancial
pe o mance, making i a mo e appealing in es men . Aligning wi h ESG p inciples can help
i ms imp o e hei ma ke posi ioning, boos in es o con idence, and ensu e long- e m
success.
Kolbel e al. (2020) a gue ha in es o s who wish o ha e a genuine impac should
collabo a e wi h hei po olio i ms and o he in es o s o inc ease hei in luence. They eel
ha addi ional policies, such as pollu ion le ies, s ic e en i onmen al ules, and inancial
incen i es, a e equi ed o achie e meaning ul p og ess. These egula ions may imp o e he
ma e ial inancial bene i s o ESG p ac ices, making sus ainable business s a egies mo e
economically iable and appealing o bo h in es o s and i ms. Fu he mo e, Lingnau e al.
(2022) show ha business sus ainabili y has an unequal impac on in es o beha iou .
Al hough good sus ainabili y ope a ions may no necessa ily esul in an inc ease in WTI,
ailu e o achie e undamen al sus ainabili y c i e ia se e ely educes in es o in e es . This
inding sugges s ha , o p i a e in es o s, he ques ion is no so much whe he "i pays o be
good", bu whe he "i hu s o be bad". Companies ha iola e sus ainabili y s anda ds ace
penal ies, and nega i e epu a ional damage ou weighs any po en ial o inc eased inancial
e u ns. The impac o hese indings highligh s he g owing impo ance o ESG a iables in
in luencing in es o beha iou and equi y in es men decisions.
Bahoo e al. (2024) ocus in hei pape on AI’s impac in a ious aspec s o inancial sys ems,
such as ma ke p edic ion, ola ili y educ ion, isk mi iga ion, inancial s abili y,
pe o mance e alua ion, aud de ec ion, and ea ly wa ning models o c isis p e en ion. The
s udy highligh s he widesp ead adop ion o AI in inance, u ging i ms o emb ace hese
echnologies o emain compe i i e. Policymake s a e encou aged o suppo AI adop ion
h ough unding and aining ini ia i es. The au ho s also acknowledge he s udy’s
limi a ions, including he b oad scope o opics co e ed and he e ol ing na u e o
AE
AI Cha bo s: Fas T acking Sus ainabili y Repo Analysis
o Enhanced Decision Making
1246 Am i ea u Economic
echnological ad ancemen s while men ioning ha u he esea ch di ec ions should del e
deepe in o speci ic subjec s and explo e he implica ions o ecen AI- ela ed de elopmen s
o a ious domains.
The wo k o Aga wal e al. (2022) ocused on a bibliome ic analysis o he mos ci ed
publica ions on cha bo s and i ual assis an s, in he con ex o he dynamics o echnological
p og ess. The au ho s highligh he ac ha he USA has he la ges scien i ic p oduc ion on
cha bo s, he end o publica ions being con inuously inc easing in ecen yea s.
In he a icle by Pillai and Si a hanu (2020), he beha iou al in en ion o Indian cus ome s
and he ac ual use o AI-powe ed cha bo s in he ou ism and hospi ali y sec o is analysed.
In he s udy, he in e iew echnique was used and hen he collec ed in o ma ion was
analysed wi h NVi o 8.0. The esul s o he s udy e lec he clea in en ion o use he cha bo ,
he use ulness, and he us placed in i by he po en ial use s.
The ypes o ela ionships ha consume s o p oduc s and se ices ha e wi h cha bo s c ea ed
by a i icial in elligence we e esea ched by Youn and Jin (2021). The au ho s subs an ia ed
ha , in he case o he compe en pe sonal b and, consume s pe cei e an assis an cha bo as
mo e compe en han a iendly cha bo .
The use o a i icial in elligence o cos con ol in he condi ions o he ansi ion om a
linea o a ci cula economy model is he subjec o esea ch by he au ho s Zo a, Cîmpeanu
and D agomi (2023). The au ho s p esen ed and analysed i e p oposed ci cula economy
cha bo s, also p o iding hei de elopmen and es ing p ocedu es using na u al language
p ocessing and deep p ocessing echniques. I is concluded ha he in eg a ion o a i icial
in elligence in he ci cula economy can ensu e he educ ion o was e and pollu ion, he
sa ing o companies' esou ces, and he inc ease o hei pe o mance.
Abdulquad i e al. (2021) analysed h ough he Sea ch-Access-Tes model he ole o
cha bo s used by banks in Nige ia in changing he business model and inc easing cus ome
engagemen and hei access o inance. The esul s o he au ho s' s udy show ha cha bo s
on Wha sApp a e used by mos banks in Nige ia; he language used was only English and
he cha bo s p esen ed hemsel es wi h a emale gende iden i ica ion.
Noy and Zhang’s (2023) pape in es iga es how he Cha GPT AI cha bo a ec s he
p oduc i i y o p o essional w i e s. The esul s e eal signi ican inc eases in p oduc i i y
by educing ask comple ion ime whils imp o ing ou pu quali y. The esul s also indica e
a dec ease in inequali y be ween wo ke s, as lowe -skilled wo ke s bene i he mos om
using he cha bo . The esul s indica e ha he cha bo is mo e e icien subs i u ing wo ke
e o , a he han complemen ing hei skills. Pa icipan s exposed o he cha bo also
epo ed inc eased job sa is ac ion and sel -e icacy, along wi h exci emen and conce ns
abou he echnology. The au ho s p opose u u e esea ch a ge ed in unde s anding
Cha GPT’s b oade impac and o add ess he conce ns ha people ha e ega ding i s
po en ial o dis up labou ma ke s.
Odonko e al.’s (2024) s udy, wi h a ocus on inancial epo ing, audi ing and decision
making, i is explo ing how AI is anso ming accoun ing p ac ices. The au ho s e iew
di e en scien i ic li e a u e and case s udies om he pas decade o unde s and AI’s impac ,
New T ends in Sus ainable Business and Consump ion
AE
Vol. 26 • Special Issue No 18 • No embe 2024 1247
e ec i enes and he challenges aced in accoun ing. Thei indings indica e ha AI
signi ican ly imp o es accu acy and e iciency in inancial epo ing by au oma ing asks and
enabling p edic i e analy ics. Bu hese imp o emen s do no come wi hou a cos , as
challenges such as he need o skilled employees and da a p i acy conce ns exis . A balanced
app oach ha emphasises con inuous lea ning, and e hical conside a ions o AI in eg a ion
in accoun ing is p oposed by he au ho s.
2. Ways o build and es a cha bo o acili a e he analysis o inancial documen s
Wi h he gain in popula i y expe ienced by La ge Language Models (LLMs) wi h he ad en
o he publicly a ailable e sion o Open AI’s Cha GPT, se e al s a -ups and companies
s a ed weaking he unde lying models o building on op o hem o c ea e cus om solu ions
ha excel o speci ic asks. As people a e ying o inc ease hei p oduc i i y o make hei
wo k easie , se e al hi d-pa ies a e o e ing solu ions ha help use s achie e jus ha . One
such solu ion is p esen ed in his pape – Cha base.
Cha base (h ps://www.cha base.co/) is a web pla o m ha le e ages Open AI’s Gene a i e
P e- ained T ans o me (GPT) LLM in o de o build cus om A i icial In elligence (AI)
cha bo s in mo e han 80 languages.
The s eps o building a cus om GPT cha bo using he Cha base pla o m a e qui e
s aigh o wa d:
1. Impo ing he da a.
I he AI cha bo is mean o only o pe sonal use, o any o he cus omisa ions a e no
needed, he p ocess can be s opped he e. O he wise, he nex s eps se e o enhance he
cha bo and use -expe ience.
2. Cus omising he beha iou and appea ance o he cha bo .
3. Embedding he cha bo on a websi e.
4. In eg a ing he cha bo wi h se e al ools.
Fo he i s s ep: impo ing he da a, se e al da a sou ces, such as .pd , .doc, .docx, . x iles
o webpages o be c awled, can be added in o de o he model o be ained on i and p o ide
he p ope answe s and sou ces o hose answe s. Fo he ee e sions o he pla o m, he
da a is limi ed o 400.000 cha ac e s pe cha bo , o oughly a 5MB ile, while he numbe o
links is limi ed o 10.
In o de o demons a e he capabili ies o he AI cha bo , in his pape , we ha e chosen
Romgaz’s sus ainabili y epo o he yea 2023 (a ailable he e:
h ps://www. omgaz. o/en/2023-sus ainabili y- epo ). Wi h 131 pages and 289,103
cha ac e s, Romgaz’s sus ainabili y epo can p o ide us wi h enough in o ma ion needed
o assess he solu ion.
AE
AI Cha bo s: Fas T acking Sus ainabili y Repo Analysis
o Enhanced Decision Making
1248 Am i ea u Economic
A company's sus ainabili y epo gi es a de ailed analysis o how he company inco po a es
en i onmen al, social, and go e nance (ESG) p inciples in o i s business s a egy and
ope a ions, demons a ing how sus ainabili y is linked o long- e m alue gene a ion. I
includes majo isks and oppo uni ies, such as how he company handles clima e- ela ed
isks, egula o y changes, and eme ging ma ke ends, as well as iden i ying inno a ion
oppo uni ies and long- e m sus ainable g ow h. En i onmen al pe o mance and impac a e
add essed using da a on emissions, ene gy use, wa e managemen , and was e educ ion, as
well as ini ia i es aimed o educe ecological oo p in s. Employee well-being, di e si y,
human igh s, in ol emen in he communi y, e hical beha iou , and boa d di e si y a e all
examples o social and go e nance conce ns. The epo also con ains de ailed pe o mance
me ics and da a, such as impo an ESG indica o s and a ge s, which a e equen ly e i ied
by hi d pa ies, o e ing anspa ency and accoun abili y o sus ainable p og ess.
Fo he second s ep: cus omising he beha iou and appea ance o he cha bo , se e al se ings
can be adjus ed o cus omise he cha bo ’s ou pu and appea ance. The i s and a guably
mos impo an se ing o be cus omised is he “Ins uc ions” ield. The “Ins uc ions” ield
con ains he model’s sys em p omp . The sys em p omp se es as gua d ails ha ell he
unde lying model how o beha e. E e y use ’s inpu will be amended o he sys em p omp
be o e i is passed o he unde lying GPT model.
Fo his pape , in o de o ensu e eplicabili y and an e alua ion o he mos basic model, we
ha e chosen o use he de aul sys em p omp ha s a ed: “I wan you o ac as a suppo
agen . You name is "AI Assis an ". You will p o ide me wi h answe s om he gi en in o.
I he answe is no included, say exac ly "Hmm, I am no su e." and s op a e ha . Re use
o answe any ques ion no abou he in o ma ion. Ne e b eak cha ac e .”
The a o emen ioned sys em p omp s a s wi h desc ibing he ole ha he unde lying model
should assume and hos i should beha e – as a suppo agen . Then, i ells he model how i
should espond – p o iding answe s om he gi en in o (i.e. he p o ided documen s). Then,
he limi a ions o he model a e included – when he in o ma ion is no p esen in he da a
p o ided, he cha bo should say “Hmm, I am no su e.”. To u he ein o ce his beha iou ,
and a oid any hallucina ions om he model, he sys em p omp ells he model o e use o
answe any ques ion ha is no in he in o ma ion p o ided. Hallucina ions e e o he
endency o Na u al Language P ocesso s o p o ide ou pu ha con ains undesi ed con en
o non-sensical ma e ial ha de ia es om he sou ce ma e ial (Ziwei e al., 2023). Finally,
he model is eminded o ne e b eak cha ac e du ing he in e ac ion. O cou se, he e a e
se e al ways in which his sys em p omp can be cus omised o be e i he needs o
someone who needs o ex ac in o ma ion om he inancial documen s p o ided, bu
p omp op imisa ion is no wi hin he scope o his pape .
The nex cus omisa ion s ep a ailable is he model selec ion. Cha base suppo s he ollowing
Open AI models: gp -3.5- u bo, gp -4- u bo, and gp -4. Fo his pape , we ha e chosen he
eely a ailable gp -3.5- u bo model in o de o assess he cha bo s’ de aul capabili ies. I is
also impo an o no e ha he gp -3.5- u bo model is bo h he mos en i onmen ally and cos -
e ec i e selec ion o his cha bo , o e ing 90% educed cos s when compa ed o i s gp -4-
u bo coun e pa and 95% educed cos s when compa ed o he gp -4 model.
New T ends in Sus ainable Business and Consump ion
AE
Vol. 26 • Special Issue No 18 • No embe 2024 1255
E ns , E., Me ola, R. and Samaan, D., 2018. The economics o a i icial in elligence:
Implica ions o he u u e o wo k. [online] h ps://doi.o g/10.13140/
RG.2.2.29802.57283.
IMF, 2021. Powe ing he Digi al Economy, [online] A ailable a :
<h ps://www.elib a y.im .o g/downloadpd /jou nals/087/2021/024/a icle-A001-
en.xml> [Accessed 1 Ap il 2024].
Ji, Z., Lee, N., F ieske, R., Yu, T., Su, D., Xu, Y., Ishii, E., Bang, Y., Chen, D., Dai, W.,
Chan, H.S., Mado o, A. and Fung, P., 2022. Su ey o Hallucina ion in Na u al Language
Gene a ion. [online] h ps://doi.o g/10.48550/ARXIV.2202.03629.
Kölbel, J.F., Heeb, F., Pae zold, F. and Busch, T., 2020. Can Sus ainable In es ing Sa e he
Wo ld? Re iewing he Mechanisms o In es o Impac . O ganiza ion & En i onmen ,
33(4), pp. 554-574. h ps://doi.o g/10.1177/1086026620919202.
Lin, W.Y., 2023. P o o yping a Cha bo o Si e Manage s Using Building In o ma ion
Modeling (BIM) and Na u al Language Unde s anding (NLU) Techniques. Senso s,
23(6), p. 2942. h ps://doi.o g/10.3390/s23062942.
Lingnau, V., Fuchs, F. and Beham, F., 2022. The link be ween co po a e sus ainabili y and
willingness o in es : new e idence om he ield o e hical in es men s. Jou nal o
Managemen Con ol, 33(3), pp. 335-369. h ps://doi.o g/10.1007/s00187-022-00340-z.
Mehwish, Mus a a, A. And Shahid, J.K., 2022. ESG Fac o s and hei In luence on he
In es men Beha io o Indi idual In es o : A Case om Pakis an. In e na ional Jou nal
o Business and Managemen Sciences, 3(3), pp. 21-45.
Noy, S. and Zhang, W., 2023. Expe imen al E idence on he P oduc i i y E ec s o
Gene a i e A i icial In elligence. SSRN Elec onic Jou nal. [online] h ps://doi.o g/
10.2139/ss n.4375283.
Odonko , B, Kaggwa S., Uwaoma,P., U., Hassan A.O., and Fa ayola O.A., 2024. The impac
o AI on accoun ing p ac ices: A e iew: Explo ing how a i icial in elligence is
ans o ming adi ional accoun ing me hods and inancial epo ing. Wo ld Jou nal o
Ad anced Resea ch and Re iews, 21(1), pp. 172-188. h ps://doi.o g/10.30574/
wja .2024.21.1.2721.
OpenAI, 2024a. GPT-3.5 Tu bo ine- uning and API upda es. [online] A ailable a :
<h ps://openai.com/blog/gp -3-5- u bo- ine- uning-and-api-upda es>
OpenAI, 2024b. Pla o m Open AI. [online] A ailable a : <h ps://pla o m.openai.com/
docs/models/o e iew> [Accessed 1 Ap il 2024].
Pa k, S.R. and Jang, J.Y., 2021. The Impac o ESG Managemen on In es men Decision:
Ins i u ional In es o s’ Pe cep ions o Coun y-Speci ic ESG C i e ia. In e na ional
Jou nal o Financial S udies9(3), p.48. h ps://doi.o g/10.3390/ij s9030048.
Pillai, R. and Si a hanu, B., 2020. Adop ion o AI-based cha bo s o hospi ali y and ou ism.
In e na ional Jou nal o Con empo a y Hospi ali y Managemen , 32(10), pp. 3199-3226.
h ps://doi.o g/10.1108/IJCHM-04-2020-0259.
Sul ana, S., Zulki li, N. and Zainal, D., 2018. En i onmen al, Social and Go e nance (ESG)
and In es men Decision in Bangladesh. Sus ainabili y, 10(6), p. 1831.
h ps://doi.o g/10.3390/su10061831.
AE
AI Cha bo s: Fas T acking Sus ainabili y Repo Analysis
o Enhanced Decision Making
1256 Am i ea u Economic
Thomas, R., Amico, R. and Kolbjø ns ud, V., 2016. How a i icial in elligence will ede ine
managemen . Ha a d Business Re iew, [online] A ailable a : <h ps://s o e.hb .o g/
p oduc /howa i icial-in elligence-will- ede ine-managemen /h0380z?sku=H0380Z-
PDF-ENG> [Accessed 1 Ap il 2024].
Wamba-Taguimdje, S.-L., Fosso Wamba, S., Kala Kamdjoug, J.R. and Tcha chouang
Wanko, C.E., 2020. In luence o a i icial in elligence (AI) on i m pe o mance: he
business alue o AI-based ans o ma ion p ojec s. Business P ocess Managemen
Jou nal, 26(7), pp. 1893-1924. h ps://doi.o g/10.1108/BPMJ-10-2019-0411.
Youn, S. and Jin, S.V., 2021. “In A.I. we us ?” The e ec s o pa asocial in e ac ion and
echnopian e sus luddi e ideological iews on cha bo -based cus ome ela ionship
managemen in he eme ging “ eeling economy”. Compu e s in Human Beha io , 119,
a . no. 106721. h ps://doi.o g/10.1016/j.chb.2021.106721.
Zo a, R.D., Cîmpeanu, I.A. and D agomi , D.A., 2023. Use and Design o Cha bo s o he
Ci cula Economy. Senso s, 23(11), a . no. 4990. h ps://doi.o g/10.3390/s23114990.
