Introducción

The institutionalization of artificial intelligence (AI) is accompanied by a fragmentation of the regulatory landscape, an asymmetry in the legal efficacy of various legislative acts (such as the AI Act, GDPR, ISO/IEC 42001), and a conspicuous absence of a cohesive model for implementing the principles of fairness, transparency, privacy, accountability, and auditability. The lack of formalized mechanisms for metric evaluation and procedural compliance necessitates the development of a comprehensive framework that integrates ethical, technical, and legal requirements into a unified system of certification, monitoring, and accountability for high-risk AI systems (^{Riabokon, 2025}).

The purpose of this study is to formulate and substantiate an optimized ethical and legal framework for the regulation of artificial intelligence applications (SCALEs-15), which ensures the integration of the principles of fairness, privacy, transparency, accountability, and auditability through normatively oriented implementation mechanisms, metric verification, and alignment with international regulatory acts (AI Act, GDPR, ISO/IEC 42001).

To achieve the stated purpose, the following tasks were undertaken:

Literature Review

Within the framework of the current study, we undertook a comprehensive review of the latest legal scholarship pertaining to ethical norms and the regulation of artificial intelligence (AI) utilization.

In particular, ^{Ashraf and Mustafa (2025)} substantiated the imperative for regulatory convergence of ethical and legal standards governing AI, particularly concerning the assurance of algorithmic fairness, explainability, privacy by design and legal accountability. The researchers discerned a significant disparity between technical regulation and legal support, which complicates the implementation of ethical AI principles across transnational jurisdictions.

This perspective is shared by ^{More et al. (2025)}, who examined the normative and ethical challenges associated with AI implementation within the legal framework, concentrating on the risks of algorithmic bias, opacity, accountability deficits, and privacy risks. The authors emphasized the need for an ethical and legal balance between technological innovation and social responsibility, articulating recommendations for the establishment of a sustainable and trustworthy regulatory environment for AI in jurisprudence.

In turn, ^{Lund et al. (2025)} conducted a comparative legal analysis of transparency frameworks (IEEE P7001, CLeAR), elucidating the progression of ethical and legal paradigms aimed at ensuring AI іnterpretability across diverse legal frameworks. Their contribution highlight underscore the necessity for differentiated regulatory approaches to transparency, considering the associated risks, the balance between transparency and the protection of intellectual property, as well as the formalization of the obligation for ongoing documentation as a vehicle for legal accountability.

This issue was thoroughly examined by ^{Nabil et al. (2025)}, who investigated the ethical and legal risks of using AI in Information Technology project management, with a particular focus on algorithmic bias, data privacy, and deficitary governance. The study emphasizes the necessity of developing transparent and accountable regulatory frameworks that ensure compliance with the General Data Protection Regulation (GDPR), the California Consumer Privacy Act (CCPA), and the Health Insurance Portability and Accountability Act (HIPAA), while simultaneously reducing the risk of algorithm-driven bias and discrimination in managerial decision-making.

A domain-specific analysis by ^{Mishra et al. (2025)} examined the ethical and legal risks arising from the use of artificial intelligence and machine learning (AI/ML) systems in credit scoring models for Buy Now, Pay Later (BNPL) services in the United States, with particular attention to violations of the Equal Credit Opportunity Act (ECOA), the Fair Credit Reporting Act (FCRA), and established fair lending principles.

Their contribution highlights for the modernization of regulatory practices to eliminate algorithmic bias, ensure explainability, and safeguard the rights of vulnerable groups within the domain of AI regulation in finance.

An additional industry-specific perspective was considered by ^{Janjua et al. (2025)}, who probed the deployment of AI-driven cloud services in Human Resource Management in Pakistan, highlighting significant ethical and legal risks: algorithmic bias, opacity, and violations of data protection. The researchers highlighted the pivotal role of ethical leadership in fostering fairness, accountability, and legal compliance while adapting AI to the national cultural and legal context.

Next, ^{Sonani and Govindarajan (2025)} introduced an innovative cloud-based model for normative and ethical reinforcement, which amalgamates multi-level compliance, explainable artificial intelligence (XAI), as well as reward algorithms to ensure transparency, fairness, and legal compliance within AI systems under regulatory oversight. This line of research contends that, privacy-by-design, and cross-jurisdictional legal adherence are imperative prerequisites for the ethically sound deployment of AI technologies in law enforcement.

In another significant contribution, ^{Ferhataj et al. (2025)} empirically validated the heightened ethical and legal sensitivity among students in terms of utilizing the AI and robotics, particularly in the areas of transparency, accountability, and data privacy. The study introduces the Ethical Awareness-Trust Framework model, which integrates ethical literacy, institutional trust, and regulatory advocacy as essential conditions for the responsible implementation of AI in accordance with the stipulations of the European Union AI Act and European civil law pertaining to robotics.

Further, ^{Stamboliev and Christiaens (2025)} scrutinized the concept of Trustworthy AI as a regulatory signifier, arguing that the European Guidelines on Trustworthy AI (EGTAI) represent only a superficial consensus among industry stakeholders, ethicists, and legal experts. This analysis demonstrates that that the discursive dominance of Big Tech replaces ethical imperatives with economic motivations, thereby undermining the autonomy of AI ethics within legal regulation.

Lastly, the research problem was summarized by ^{Poli et al. (2025)}, who conducted a critical literature review alongside a case analysis of ethical dilemmas in AI, focusing on issues of fairness, transparency, accountability, and algorithmic bias. The scholars pinpointed a lack of research in the practical dimension of AI ethics and underscored the necessity of establishing a comprehensive interdisciplinary research agenda to foster the ethical and legal advancement of technological innovations.

The examination of contemporary scholarship has revealed the fragmentation inherent in AI ethics and legal regulation, highlighting a significant deficiency of effective mechanisms for fairness, іnterpretability, accountability, and data protection. Moreover, the lack of alignment between sector-based and transnational regulatory regimes reflects broader gaps in coordination and harmonization across AI governance structures. The preeminence of industrial interests over ethical paradigms was underscored, alongside a call for the establishment of a cohesive, and adaptive framework that ensures multilevel compliance, dynamic accountability, and legal legitimacy within high-risk sectors.

Methods and Materials

Research procedure

The study was conducted according to the following scheme (Figure 1).

Figure 1 Research procedure 

Methods

The study adopted a sequential multi-stage approach, where every methodological approach has been utilized at distinct stages of the research process to guarantee methodological cohesion and scientific transparency. In the first stage of the analytical phase, a comparative ethical-legal analysis was used in evaluating principal normative and regulatory instruments on artificial intelligence such as for example the AI Act, General Data Protection Regulation and ISO/IEC 42001. This approach facilitated a consistent contrast in respect of legal authority; ethical reach; and enforcement and identified regulatory differences, normative lacunae, and conflicting governance approaches. At the next stage, the normative model decomposition was performed for the SCALEs-15 framework. In this phase, the framework’s principles were systematically classified into legal, technical, and procedural categories. They were assessed for internal consistency, component traceability, and compatibility with the existing multi-act regulatory environment. A metric conformity assessment technique was introduced in the third phase to operationalize ethical and legal principles. The degree to which the frameworks discussed were embedded with fairness, privacy, transparency, accountability and auditability was assessed quantitatively and semi-quantitatively on indicators such as Demographic Parity, Privacy Risk Score, Explainability Index. Next, another refined version of the SCALEs-15 was created through normative modeling. This process integrated regulatory requirements, technical standards, and ethical metrics into an operational framework that was unified and operationally consistent. Legal forecasting was then applied in the last phase of the evaluation process, to evaluate the prospects of regulatory legitimation of the optimized SCALEs-15 Framework. Such an assessment was done in the context of European Union law and international ISO/IEC standards, emphasizing certification paths, audit schemes and procedures of compliance. The study was conducted between January 2024 and March 2025.

Tools

To evaluate the regulatory frameworks governing the utilization of artificial intelligence, relevant ethical and legal metrics were employed, which were subsequently categorized in accordance with the principles of AI ethics and governance (Table 2).

Each metric (Table 2) serves as an instrumentalization of an abstract ethical principle. Accordingly, this distribution facilitates the formalization of AI audits, the construction of comprehensive checklists, as well as the verification of AI alignment with regulatory and ethical standards.

Results

In accordance with the research framework, a comprehensive ethical and legal comparative analysis was conducted on the regulatory acts (frameworks) regarding the implementation of artificial intelligence (Table 1). This analysis was executed in alignment with the fundamental principles of AI ethics & governance, consisting of relevant metrics (Table 2) ‒ Table 3.

The ethical and legal analysis (

Table 3) has elucidated that the AI Act (EU) emerges as the most balanced regulatory framework, mandating certification for high-risk AI applications. While GDPR, PIPL, and ISO/IEC 42001:2023 offer a substantial level of data governance, they only partially address the critical dimensions of fairness and explainability. Soft law instruments such as ALTAI, EGTAI, and IEEE P7001 advocate for ethical compliance but lack binding authority. The ECOA and FCRA continue to hold relevance in the financial sector; however, they are inadequately equipped to confront the challenges posed by AI. Metrics like Demographic Parity and Explainability Index allow for formalizing the assessment of alignment with ethical and legal principles.

A comparative analysis of existing ethical and legal frameworks concerning AI revealed the fragmentation within the regulatory landscape, the absence of a cohesive compliance model, and an imbalance between legal and technical mechanisms. In this light, the implementation of the SCALEs-15 framework (Standards for Certified AI & Legal Ethics System ( 15 Principles) is warranted as a comprehensive system for certification, auditing, and metric evaluation, thereby ensuring the integration of the principles of fairness, privacy, transparency, accountability, and human agency. Decomposition analysis enables the verification of internal structural consistency, the resolution of regulatory conflicts, and the adaptation of SCALEs-15 to transjurisdictional challenges ‒ Table 4.

Decomposition analysis (Table 4) elucidated the conceptual integrity of SCALEs-15, yet unveiled regulatory discrepancies that pose a risk of ethical and legal conflicts. The disjunctions between transnational norms, the absence of appeal mechanisms, the partial implementation of the principles of fairness and accountability, and the threat of formalized consent (dark patterns) have been identified. This underscores the imperative to optimize the framework to harmonize alignment criteria and facilitate transjurisdictional integration ‒ Table 5.

Based on the developed optimization solutions (Table 5), it has been ascertained that the refined SCALEs-15 framework exhibits a substantial degree of conceptual consistency with the key ethical and legal stipulations governing AI regulation. Nonetheless, the practical implementation of each principle requires the elucidation of assessment tools tailored to regulatory acts (AI Act, GDPR, ISO/IEC 42001, OECD Principles, etc.). The proposed optimization solutions are pertinent to current challenges; however, their efficacy must be further substantiated through a re-evaluation utilizing specialized metrics (Table 2) compared to the previous version of SCALEs‑15 and the most balanced regulatory act currently in effect, the AI Act (Table 6).

The conclusive findings indicate that the AI Act (EU) is the most legally robust framework; however, it necessitates instrumental detailing. The basic SCALEs-15 is primarily declarative, whereas the optimized version ensures maximal adherence to essential principles through the integration of metric indicators and procedural verification. It is expedient to regard this as a promising paradigm for the ethical-legal standardization of artificial intelligence, warranting further regulatory legitimization.

Discussion

Let's verify the obtained results for alignment with analogous studies within a specified chronological period. The study conducted by ^{Radanliev (2025)} documented the fragmentation of global ethical and legal regulation of artificial intelligence and proposed technical compensatory measures. The investigation substantiates that the optimized SCALEs-15 facilitates a systematic implementation of the principles of fairness, privacy, and transparency through normatively-oriented mechanisms, thereby establishing a foundation for legal harmonization.

^{Agbadamasi et al. (2025a)} highlighted a regulatory deficiency in the USA: fragmented norms (FTC Act, Algorithmic Accountability Act, HIPAA) inadequately address algorithmic opacity, bias, and data surveillance. In contrast, this study demonstrates that the optimized SCALEs-15 demonstrates the metric integration of fairness, accountability, and privacy-by-design, ensuring normative consistency and the potential for transjurisdictional integration.

^{Zahra (2025)} documented a critical discrepancy between the operational efficacy of AI within jurisprudence and the lack of ethical and legal guarantees. Instead, this study elucidated a normative synergy within SCALEs-15: the integration of explainable AI (XAI), fairness audits, and traceability emerges as indispensable mechanisms for legal legitimation.

^{Bahangulu and Owusu-Berko (2025)} underscored the corporate risks associated with algorithmic bias and the inadequacies of data ethics in business analytics, highlighting the weakness of self-regulation. In contrast, this study’s optimized SCALEs-15 provided a robust institutional framework for fairness-aware machine learning, incorporating layers of explainability and normative audits, thereby transforming data ethics into a legally binding component of corporate AI governance.

^{Kamatala et al. (2025)} concentrated on algorithmic bias and proposed a technically oriented debiasing model without normative integration. Notably, the current research advanced SCALEs-15 as a multi-layered regulatory architecture, wherein fairness constraints, XAI, and legal accountability are seamlessly woven as mandatory elements for the legal legitimation of AI decision-making.

^{Vujnovic et al. (2025)} accentuated the role of PR specialists as agents of public explainability, yet neglected the normative imperative of transparency. In contrast, the current study’s optimized SCALEs-15 established transparency as a legal requirement through metrics of explainability, auditability, and traceability, thereby transforming communicative transparency into a regulatory obligation.

^{Agbadamasi et al. (2025b)} found that the implementation of corporate AI governance in the USA predominantly transpires within internal business intelligence systems, characterized by a dominance of compliance-oriented control mechanisms. In contrast, the current study has formulated and refined SCALEs-15 as a normatively structured framework that transcends conventional corporate practices, while ensuring the formalized integration of fairness, accountability, and transparency metrics with legal validation and procedural accountability.

^{Al-Harbi (2025)} elucidated that the financial application of AI is marked by regulatory archaism, particularly concerning bias modules, weak explainability, and the lack of sector-specific regulatory stratification. In the present study, the aforementioned issues have been transformed into an operational-legal framework - the optimized SCALEs-15 - which has facilitated standardized verification of fairness, the legal validity of decisions, and transparency through built-in control metrics and procedural alignment with regulatory acts.

^{Ramesh (2025)} accentuated the ethical risks associated with AI/ML in the insurance sector, particularly bias and the lack in transparency. This investigation has normatively delineated the said challenges through the optimized SCALEs-15, which incorporates built-in fairness audits, explainability modules, and legal accountability in the area of insurance scoring.

Ultimately, ^{Alhasan (2025)} found that the integration of AI into arbitration engenders conflicts with the principles of procedural fairness, legal accountability, and the legitimacy of decisions, particularly regarding the inadequacy of the New York Convention for AI-generated awards. In the current study, these deficiencies have been rectified by optimizing SCALEs-15, wherein the principles of explainability, reversibility, and accountability have been institutionalized as prerequisites for the legal legitimacy of AI arbitration.

The analysis of scholarly publications corroborated the fragmentation of the AI regulatory field and the scarcity of ethical and legal guarantees. This study has substantiated that the optimized SCALEs-15 engenders a cohesive regulatory architecture with integrated modules for fairness, privacy, explainability, accountability, and auditability, thereby positioning it as a prospective transjurisdictional standard.

Conclusions

Drawing upon a comprehensive ethical and legal examination of both existing and proposed AI regulatory frameworks, utilizing analytical metrics for fairness, privacy, explainability, accountability, auditability, and human agency, it has been determined that the regulatory landscape is notably fragmented. Most current legislative instruments, including the AI Act, GDPR, and ISO/IEC 42001, exhibit only partial or sector-specific implementations of the relevant principles. At the same time, the proposed optimized SCALEs-15 model is distinguished by maximum conceptual completeness, procedural organization, and the integration of metric indicators, thereby ensuring a systemic verification of AI decision-making in accordance with ethical and legal norms.

A comparative analysis employing specialized metrics showed that the optimized SCALEs-15 outperforms the AI Act in its elaboration of operational mechanisms and effectively addresses the shortcomings of existing regulatory frameworks. In light of this, it is expedient to advocate for the legal endorsement of the optimized SCALEs-15 as a cohesive framework for the ethical and legal governance of AI, with subsequent harmonization to European and international standards (AI Act, GDPR, ISO/IEC 42001).

The scientific novelty of this research lies in the formalization and decomposition of the optimized SCALEs-15 framework, conceived as a multilevel normative model for the ethical and legal verification of artificial intelligence. This framework, for the first time, integrates metric indicators into the mechanisms of procedural compliance.

The practical significance of the findings resides in the establishment of a unified structure for normative auditing, certification, and risk management of AI decisions. This framework can be utilized by developers, auditors, and regulators in the process of assessing the compliance of artificial intelligence with existing ethical and legal frameworks.