Revista Multidisciplinaria Perspectivas Investigativas

Multidisciplinary Journal Investigative Perspectives

Vol. 6(1), 47-57, 2026

https://doi.org/10.62574/rmpi.v6i1.507

Strict liability for damages arising from artificial intelligence: A critical

analysis in light of the risk framework of the Ecuadorian Civil Code

https://orcid.org/0009-0002-1286-9494

Rously Eedyah Atencio-González

reatenciog@ube.edu.ec

Universidad Bolivariana del Ecuador, Durán, Guayas, Ecuador

https://orcid.org/0000-0001-6845-1631

Geovanna Michelle Nájera-Tello

Red de Investigación Koinonia, Guayaquil, Guayas, Ecuador

https://orcid.org/0000-0003-3241-6740

ABSTRACT

The aim of the research was to analyse strict civil liability for damages arising from artificial intelligence in

relation to the risk framework of the Ecuadorian Civil Code. A meta-analysis design was adopted, involving

a systematic search of the Scopus, Web of Science and Latindex databases, covering the period from

January 2015 to March 2026, and including 27 sources that met the established selection criteria. Five

recurring legal phenomena were identified: diffuse causation, subjective fragmentation of liability,

informational asymmetry, obsolescence of the concept of fault, and a regulatory coverage gap, which

La investigación tuvo como objetivo analizar la responsabilidad civil objetiva por daños derivados de la

inteligencia artificial frente al sistema de riesgos del Código Civil ecuatoriano. Se adoptó un diseño de

metaanálisis mediante la búsqueda sistemática en bases de datos Scopus, Web of Science, Latindex, con

un período de cobertura entre enero de 2015 y marzo de 2026, incluyendo 27 fuentes que cumplieron los

criterios de selección establecidos. Se identificaron cinco fenómenos jurídicos recurrentes: causalidad

difusa, fragmentación subjetiva de la responsabilidad, asimetría informacional, obsolescencia del concepto

de culpa y laguna normativa de cobertura, que evidencian la insuficiencia del marco civil vigente para

atribuir responsabilidad por daños algorítmicos. Se concluye que el Ecuador requiere una reforma

normativa que incorpore responsabilidad objetiva agravada, solidaridad entre actores tecnológicos,

inversión de la carga probatoria, un fondo de garantía subsidiario y obligaciones de transparencia

Vol. 6(1), 47-57, 2026

La responsabilidad civil objetiva por daños derivados de la Inteligencia Artificial: Un análisis crítico frente al sistema de riesgos

del Código Civil ecuatoriano

Strict liability for damages arising from artificial intelligence: A critical analysis in light of the risk framework of the Ecuadorian

The 21st century has established artificial intelligence (AI) as one of the most significant

transformative forces in the contemporary legal system; its widespread deployment in sectors

such as healthcare, justice, finance and public administration creates scenarios in which

decisions made by algorithmic systems have direct and indirect effects on individuals’ subjective

rights, such that legal science is compelled to ask whether the traditional instruments of civil

liability, designed to regulate voluntary human conduct, are suitable when the agent causing the

harm is a machine equipped with autonomous learning capabilities.

From a historical perspective, the law has responded to each technological revolution through

the progressive adaptation of its categories: the advent of the automobile gave rise to regimes

In the Latin American context, Ecuador lacks specific legislation regulating liability arising from

the use of AI; the Ecuadorian Civil Code, which draws on the Chilean codification tradition of the

mid-19th century, structures non-contractual liability on the basis of proven fault or the

presumption of fault in certain cases involving dangerous activities, but does not provide for a

separate regime for damages caused by advanced technological systems; this omission is

particularly serious given that Ecuadorian case law lacks established precedents in this area

and that national legal scholarship is only just beginning the debate, as noted by Narváez-López

(2019) and Concha-Flores (2024).

The central question guiding this research is formulated as follows: is the strict liability regime

provided for in the Ecuadorian Civil Code sufficient to attribute liability for damages arising from

authors from the Anglo-Saxon world have proposed frameworks of strict liability without fault, as

argued by Marchisio (2021); comparative research has examined the differential treatment

offered by systems of subjective, objective and hybrid liability, as documented by Díaz and

Mitrani (2025); and European regulatory developments, particularly the European Union’s civil

liability directives, have opened a debate that transcends continental borders, within the

framework analysed by Solaiman and Malik (2025); against this backdrop, the analysis of the

Ecuadorian case provides a perspective situated within a peripheral civil law system, with

institutional and technological capacities currently in the process of consolidation.

Thus, the objective of this research is to analyse strict civil liability for damages arising from

Artificial Intelligence in relation to the risk regime of the Ecuadorian Civil Code.

significant constructs of modern private law; unlike the subjective regime, which requires proof

Vol. 6(1), 47-57, 2026

La responsabilidad civil objetiva por daños derivados de la Inteligencia Artificial: Un análisis crítico frente al sistema de riesgos

del Código Civil ecuatoriano

Strict liability for damages arising from artificial intelligence: A critical analysis in light of the risk framework of the Ecuadorian

acceptance in Latin American legal systems through general clauses on dangerous activities, as

In the Ecuadorian Civil Code, non-contractual liability is based on Article 2214, which

establishes the obligation to compensate for damage caused by a wrongful act committed with

intent or through negligence, and on Article 2229, which provides for liability for activities that,

by their nature, are capable of causing harm to third parties; however, the application of these

provisions to AI systems faces the problem of the so-called algorithmic ‘black box’, that is, the

practical impossibility of determining precisely which element of the system caused the harmful

decision and whether there was any form of fault attributable to an identified human agent, as

noted by Rodríguez-Corría and Alba-Cazales (2025) and Jaramillo-Valdivieso (2024).

of technological risk, as it falls upon whoever derives the economic benefit from the risky

activity.

The specialist literature identifies at least three categories of harm that may be caused by AI

systems: physical harm, financial harm and harm to fundamental rights such as privacy, equality

and dignity; the first category includes bodily injury or death caused by autonomous devices

such as vehicles, surgical robots or drones, the second category includes financial losses

arising from algorithmic decisions in stock markets, credit granting or insurance management,

and the third category arises when facial recognition systems, data profiling or automated

decision-making result in discrimination or infringe upon individuals’ privacy, as described by

For legal analysis, it is pertinent to distinguish between weak AI systems, which execute specific

no human agent can fully foresee or control the consequences of algorithmic action, as argued

by Erkan and Biswas (2025) and Custers et al. (2025).

et al. (2024), Zabala-Balladares et al. (2024) and Orozco-Zavala et al. (2024) have warned of

the dangers of algorithmic bias in criminal and civil proceedings, the lack of transparency in

decision-making and the absence of accountability mechanisms when the system produces

erroneous results.

The debate over the most appropriate model of liability for AI-related damages has produced

subjective liability with evidential adaptations, such as the reversal of the burden of proof or the

Vol. 6(1), 47-57, 2026

La responsabilidad civil objetiva por daños derivados de la Inteligencia Artificial: Un análisis crítico frente al sistema de riesgos

del Código Civil ecuatoriano

Strict liability for damages arising from artificial intelligence: A critical analysis in light of the risk framework of the Ecuadorian

liability solely on the causal link between the system’s operation and the damage caused, a

The third proposal, which is becoming increasingly dominant in European and Latin American

debates, opts for a hybrid regime that combines objective and subjective elements depending

on the type of system, the degree of autonomy, the sector of activity and the type of harm

caused; this alternative envisages the shared allocation of liability between the algorithm

developer, the operator deploying it and the user utilising it, assigning to each a share

proportional to their capacity to control the system, in line with the proposals of Díaz and Mitrani

(2025), Muñoz (2025) and Custers et al. (2025).

At the European level, the European Union’s AI Liability Directive and the reform of the

The Ecuadorian Civil Code incorporates the principle of liability for dangerous activities through

provisions on damage caused by animals, the collapse of buildings and industrial activities;

Ecuadorian legal doctrine has debated whether these clauses can be interpreted broadly to

include advanced technological systems as sources of risk analogous to those contemplated by

nineteenth-century legislators, and in this debate, the argument in favour of analogical

extension maintains that the basis of the rule is not the specific nature of the dangerous object,

but the general principle that whoever creates a risk must bear the harmful consequences

arising from it, a position shared by Guanoluisa Almache et al. (2021) and Zúñiga-Hurtado and

Hurtado-Macías (2023).

However, the doctrine of analogical extension faces significant technical objections: firstly, the

principle of specificity in civil torts requires that the factual basis of the rule be sufficiently

cases of AI-related harm requires highly specialised technical knowledge that the Ecuadorian

judicial system has not yet systematically developed, as noted by Kostrzewa and Nowak (2022)

and Guamán-Quinzo and Batista-Hernández (2024).

or in a complementary special law, a differentiated liability regime for the development,

deployment and use of AI systems; a position shared by Concha-Flores (2024), Parra-

Sepúlveda and Concha-Machuca (2021), and Rodríguez-Corría and Alba-Cazales (2025).

The research adopted a qualitative meta-analysis design, aimed at the systematic and critical

synthesis of published doctrinal, regulatory and jurisprudential contributions on civil liability for

analysis typical of the health sciences, did not seek the statistical aggregation of numerical data,

Sepúlveda and Concha-Machuca (2021).

Vol. 6(1), 47-57, 2026

La responsabilidad civil objetiva por daños derivados de la Inteligencia Artificial: Un análisis crítico frente al sistema de riesgos

del Código Civil ecuatoriano

Strict liability for damages arising from artificial intelligence: A critical analysis in light of the risk framework of the Ecuadorian

The literature search was conducted in the Scopus, Web of Science, Latindex, Google Scholar

The inclusion criteria applied were as follows: a) publications in peer-reviewed journals with an

impact factor or indexed in recognised databases; b) texts that specifically addressed civil

liability in relation to AI systems; c) studies examining Ecuadorian, Latin American or

literature were identified, including models of attribution, causal link, distribution of liability and

regulatory proposals; in the second, the central arguments of each source were systematised in

The meta-analysis of the reviewed literature identifies five recurring legal phenomena that

shape the landscape of civil liability for AI-related harm in the Ecuadorian context:

interaction between the different links in the chain;

the Ecuadorian Civil Code, designed to attribute liability to a specific individual, lacks

mechanisms for distributing liability amongst multiple actors with varying levels of control over

Vol. 6(1), 47-57, 2026

La responsabilidad civil objetiva por daños derivados de la Inteligencia Artificial: Un análisis crítico frente al sistema de riesgos

del Código Civil ecuatoriano

Strict liability for damages arising from artificial intelligence: A critical analysis in light of the risk framework of the Ecuadorian

computer science; the rules of evidence in Ecuador’s General Organic Code of Procedure do

The fourth phenomenon is the obsolescence of the concept of fault for damages caused by

autonomous AI; when an AI system makes a harmful decision in the exercise of its autonomous

learning capacity, there is no voluntary human behaviour to which a finding of fault can be

applied, since the system does not act with negligence or wilful misconduct, but in accordance

with its own adaptive parameters, which may lead to consequences that no human actor

foresaw or could reasonably have foreseen; the application of the concepts of slight fault, gross

personal data by algorithmic systems, does not regulate civil liability arising from material or

Based on the phenomena identified, the research formulates a regulatory proposal structured

around five pillars, intended to be incorporated into the Ecuadorian Civil Code, either through

the reform of existing provisions on non-contractual liability or through the addition of a specific

title relating to liability for high-risk technological activities.

The first pillar proposes the incorporation of a clause on aggravated strict liability for high-risk AI

systems; this clause would establish that anyone who develops, operates or uses an AI system

of high-risk systems would be referred to an updatable technical regulation, drawn up by the

competent regulatory body, in line with the approach adopted by European regulation and with

international risk classification standards.

of control over the system; where it is not possible to determine which of the actors in the

technological chain caused the damage, they would all be jointly and severally liable to the

victim, and each party’s share of liability would be determined in a subsequent recourse action,

based on criteria such as technical capacity to intervene in the system, the economic benefit

obtained and the level of information available on the risks at the time of deployment, as

Vol. 6(1), 47-57, 2026

La responsabilidad civil objetiva por daños derivados de la Inteligencia Artificial: Un análisis crítico frente al sistema de riesgos

del Código Civil ecuatoriano

Strict liability for damages arising from artificial intelligence: A critical analysis in light of the risk framework of the Ecuadorian

evidence, in line with the arguments put forward by Bottomley and Thaldar (2023) and Wang

The fourth pillar proposes the creation of a subsidiary guarantee fund for victims of AI-related

harm, covering cases where the liable party is insolvent or cannot be identified; this fund would

be financed through a system of mandatory contributions payable by developers and operators

of systems classified as high or medium risk, proportional to the turnover generated by the

activity and the risk level of the deployed system; the mechanism has precedents in the field of

road traffic accidents in several Latin American countries, so its extension to the field of AI is

legally consistent with the region’s civil law tradition.

The fifth pillar proposes incorporating an obligation of transparency and auditability as a

The results of the meta-analysis enable a productive dialogue with theory and with precedents

in comparative law; the observation of the phenomenon of diffuse causality confirms the thesis

of Čerka et al. (2015), who noted at an early stage that causality is the Gordian knot of AI

liability, a thesis complemented by the more recent contributions of Erkan and Biswas (2025),

who emphasise that causal diffusion is exacerbated as AI systems incorporate unsupervised

learning capabilities; the Ecuadorian Civil Code, structured around a notion of direct and linear

causality, does not offer adequate tools to address this reality without substantive reforms.

spaces in which no actor in the technological chain can be held liable under current rules; the

proposal for joint and several liability with rights of recourse formulated in this research takes up

innovation when accompanied by compulsory insurance mechanisms, as documented by Díaz

and Mitrani (2025).

analysed by Bottomley and Thaldar (2023) in the field of healthcare and by Wang and Zhou

conditions of the Ecuadorian context, which include more limited institutional oversight capacity,

Vol. 6(1), 47-57, 2026

La responsabilidad civil objetiva por daños derivados de la Inteligencia Artificial: Un análisis crítico frente al sistema de riesgos

del Código Civil ecuatoriano

Strict liability for damages arising from artificial intelligence: A critical analysis in light of the risk framework of the Ecuadorian

an emerging technological development sector and a population with unequal access to digital

The discussion also highlights an unresolved tension in the literature between the need to

protect victims of AI-related harm and the need not to restrict technological development

through an excessively severe liability regime; the proposal put forward in this research seeks to

articulate a balanced solution through a scheme of differentiated liability based on the system’s

risk level, the obligation of transparency as a condition for mitigation, and the creation of a

subsidiary guarantee fund. This framework translates the best practices identified in the

assess technological evidence, on the specialised training of judges in technology law, and on

the development of a legal culture of precaution in the face of innovations with high social

An analysis of strict civil liability for damages arising from artificial intelligence in relation to the

risk framework of the Ecuadorian Civil Code shows that the current regulatory framework is

structurally insufficient to attribute liability when the agent causing the damage is an algorithmic

system equipped with autonomous learning capabilities; Articles 2214 and 2229 of the Civil

Against this backdrop, the reviewed literature converges in pointing out that strict liability,

structured on the principle that whoever introduces a risk into society must bear the harmful

consequences regardless of their diligence, represents the mechanism most consistent with the

technological nature of the causative agent; its effective application in the Ecuadorian context

requires, however, that it be complemented by a regime of joint and several liability amongst the

actors in the technological chain, the reversal of the burden of proof regarding the causal link,

the creation of a subsidiary guarantee fund for victims of insolvent or unidentifiable liable

Vol. 6(1), 47-57, 2026

La responsabilidad civil objetiva por daños derivados de la Inteligencia Artificial: Un análisis crítico frente al sistema de riesgos

del Código Civil ecuatoriano

Strict liability for damages arising from artificial intelligence: A critical analysis in light of the risk framework of the Ecuadorian

systems according to their level of risk; without these enabling conditions, even the best

Čerka, P., Grigienė, J., & Sirbikytė, G. (2015). Liability for damages caused by artificial