26
h ps:// esea ch endsjou nal.com
Online a : h ps:// esea ch endsjou nal.com ISSN No: 2584-282X
Indexed Jou nal Pee Re iewed Jou nal
INTERNATIONAL JOURNAL OF TRENDS IN EMERGING RESEARCH AND DEVELOPMENT
Volume 3; Issue 5; 2025; Page No. 26-27
Recei ed: 01-07-2025
Accep ed: 06-08-2025
Published: 09-09-2025
Legal Liabili y o AI Decisions
U am Vi Singh
Associa e P o esso , Sa aswa i Vidya Mandi Law College, Bulandshah , U a P adesh, India
DOI: h ps://doi.o g/10.5281/zenodo.17085968
Co esponding Au ho : U am Vi Singh
Abs ac
A i icial In elligence (AI) is now deeply embedded in decision-making p ocesses ac oss di e se ields such as heal hca e, banking,
anspo a ion, and e en judicial sys ems. While his in eg a ion has imp o ed e iciency, i has also gene a ed unp eceden ed legal
challenges ega ding liabili y when AI-d i en decisions cause ha m. T adi ional liabili y amewo ks, which es on human aul o in en ,
a e ill-sui ed o add ess au onomous and opaque AI sys ems. This pape examines he inadequacies o exis ing doc ines, su eys
in e na ional app oaches, and e alua es new p oposals such as he Eu opean Union’s AI Liabili y Di ec i e. I u he explo es heo e ical
models, deba es on digi al pe sonhood, and he ole o co po a e go e nance. The s udy concludes ha an e ec i e legal egime o AI mus
balance inno a ion wi h accoun abili y h ough s ic liabili y mechanisms, isk-based egula ions, and enhanced o e sigh , ensu ing bo h
ic im p o ec ion and echnological ad ancemen .
Keywo ds: A i icial In elligence, Legal Liabili y, AI Go e nance, Au onomous Sys ems, S ic Liabili y, Digi al Pe sonhood, AI
Accoun abili y, Compa a i e Ju isp udence
1. In oduc ion
The accele a ed adop ion o A i icial In elligence has
shi ed c ucial decision-making om humans o machines.
This ansi ion aises p essing legal ques ions abou
accoun abili y when AI sys ems cause ha m. Con en ional
liabili y concep s- o law, con ac law, and p oduc
liabili y- ocus la gely on human ac ion o negligence.
Howe e , AI sys ems, which o en ope a e au onomously
and h ough non- anspa en p ocesses, challenge hese
doc ines. As a esul , legal schola s and policymake s a e
explo ing al e na i e amewo ks ha ensu e jus ice o
ic ims while p omo ing esponsible AI inno a ion.
2. Legal Challenges in AI Liabili y
AI decision-making is o en desc ibed as a 'black box,'
making i di icul o ace o explain ou comes. Unlike
humans, AI lacks consciousness o in en , bo h o which a e
cen al o adi ional liabili y amewo ks. Fu he mo e, he
collabo a i e ecosys em behind AI-de elope s,
manu ac u e s, deploye s, and end-use s-complica es he
alloca ion o esponsibili y. The absence o clea aul lines
leads o unce ain y in assigning liabili y when ha m occu s.
3. Compa a i e Legal Pe spec i es
▪ Uni ed S a es: Relies hea ily on p oduc liabili y and
ica ious liabili y doc ines, holding manu ac u e s and
ope a o s accoun able o AI- ela ed ha ms.
▪ Eu opean Union: Mo ing owa ds a s ic liabili y
egime, suppo ed by he AI Liabili y Di ec i e (2022)
[2] and he AI Ac , which educe he bu den o p oo o
ic ims and s eng hen anspa ency in high- isk AI
sys ems.
▪ Ge many: Conside s ex ending co po a e liabili y
p inciples o add ess AI-d i en ha ms.
▪ I an and Simila Ju isdic ions: Depend on adi ional
no ions o human and co po a e esponsibili y bu lack
explici AI-speci ic p o isions, leading o legal
ambigui y.
▪ India (eme ging pe spec i e): While India has no
dedica ed AI liabili y law, discussions unde NITI
In e na ional Jou nal o T ends in Eme ging Resea ch and De elopmen h ps:// esea ch endsjou nal.com
27
h ps:// esea ch endsjou nal.com
Aayog and consume p o ec ion amewo ks highligh
he need o ailo ed egula ion.
▪ China: Emphasizes algo i hmic accoun abili y, da a
anspa ency, and pla o m liabili y in i s AI go e nance
app oach.
These examples e eal a g adual global end owa d s ic
liabili y combined wi h isk-based egula ion, ensu ing ha
ic ims a e no bu dened wi h p o ing aul in highly
complex AI cases.
4. Legisla i e and Regula o y De elopmen s
The EU AI Liabili y Di ec i e (2022) [2] ep esen s one o
he mos ad anced egula o y amewo ks. I mode nizes
liabili y ules by lowe ing he e iden ia y bu den on ic ims,
acili a ing claims, and encou aging isk managemen in
high- isk AI sys ems. Complemen a y ins umen s such as
he AI Ac adop a isk-based app oach, imposing s ic e
compliance equi emen s o AI in sensi i e domains such
as biome ic su eillance and c i ical in as uc u e.
O he in e na ional ini ia i es, including OECD AI
P inciples (2019) and UNESCO’s AI E hics F amewo k
(2021), emphasize anspa ency, accoun abili y, and human
o e sigh , guiding s a es owa d ha monized go e nance.
5. Theo e ical App oaches o AI Liabili y
Schola s ha e p oposed a ious models:
▪ Innocen Agen Model (Halle y): AI ac s as an agen ,
bu liabili y es s wi h human p og amme s, ope a o s,
o use s.
▪ Na u al P obable Consequence Doc ine: Human
acili a o s a e accoun able o o eseeable ha ms
esul ing om AI ac ions.
▪ Digi al Pe sonhood Deba e: Some sugges g an ing AI
limi ed legal s a us o assign esponsibili ies di ec ly,
hough mos legal sys ems cu en ly a o keeping
accoun abili y human-cen e ed.
A hyb id app oach ha combines dis ibu ed accoun abili y
wi h s ic liabili y o high- isk AI applica ions appea s
mos p ac ical.
6. Co po a e Go e nance and AI Accoun abili y
As co po a ions inc easingly deploy AI in decision-making,
boa ds o di ec o s mus ensu e e hical use, isk assessmen ,
and egula o y compliance. Failu es in o e sigh may
expose co po a ions o liabili y unde iducia y and s a u o y
obliga ions. Sound go e nance amewo ks-embedding AI
e hics commi ees, algo i hmic audi s, and compliance
mechanisms-a e essen ial o mi iga ing isks.
7. Conclusion
Legal liabili y o AI emains an e ol ing domain whe e
adi ional doc ines p o e insu icien . A balanced
amewo k mus combine s ic liabili y o high- isk AI
sys ems, manda o y insu ance schemes, and egula o y
o e sigh wi h anspa ency and human accoun abili y
mechanisms. Policymake s should a oid ex emes-ei he
s i ling inno a ion h ough o e egula ion o lea ing ic ims
unp o ec ed. Ins ead, a mul i-s akeholde model, in ol ing
go e nmen s, co po a ions, and ci il socie y, is needed o
econcile inno a ion wi h jus ice and socie al us .
8. Re e ences
1. Nagda, P. Legal Liabili y and Accoun abili y in AI
Decision-Making: Challenges and Solu ions, IJIRT,
2024.
2. Eu opean Commission, AI Liabili y Di ec i e, 2022.
3. Goyal S. AI and Legal Liabili y: Who Is Responsible
When Machines Go W ong, IJLLR, 2025.
4. Chauhan, R. AI-Powe ed Decision-Making in
Co po a ions: Legal Liabili y and Go e nance, 2025.
5. Halle y, G. A i icial In elligence and Legal Liabili y,
a Xi , 2018.
6. No on Rose Fulb igh , EU AI Ac and AI Liabili y
Di ec i e, 2024.
7. Legal Liabili ies o A i icial In elligence: An
O e iew, Jou nal o Law and Social Sciences, 2024.
8. Eu opean Commission, Liabili y Rules o A i icial
In elligence, O icial Documen a ion, 2022.
C ea i e Commons (CC) License
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 4.0) license. This license pe mi s
un es ic ed use, dis ibu ion, and ep oduc ion in any
medium, p o ided he o iginal au ho and sou ce a e
c edi ed.
