The adoption of AI-enabled and cyber-physical technologies in security and defence environments introduces complex legal, ethical, safety, and operational constraints that must be evaluated consistently across the capability lifecycle. Existing mechanisms such as Article 36 reviews, safety cases, and governance assessments typically operate in isolation, lacking an integrated engineering-oriented decision model aligned with acquisition processes. This paper presents a lifecycle-integrated Ethical-Legal Assessment (ELA) framework designed as a structured, evidence-based assurance mechanism embedded within defence acquisition gates. The framework combines non-derogable legal hard stops with a quantitative multi-axis scoring model covering legality, ethical acceptability, safety and robustness, and accountability. A formal decision logic translates assessment results into GO, CONSTRAINED GO, or NO-GO outcomes, enabling traceable and auditable risk-based authorization. To validate feasibility, a Python-based prototype implementing structured checklists, rule-based aggregation, risk identification, and risk visualization was developed. The framework was evaluated in an urban UAS-ISR scenario, demonstrating how lifecycle evidence updates axis scores and supports controlled deployment decisions.
Ethical-Legal Assessment in the Adoption Process of New Defence Technologies in Space
Boi, Biagio;Esposito, Christian;
2026
Abstract
The adoption of AI-enabled and cyber-physical technologies in security and defence environments introduces complex legal, ethical, safety, and operational constraints that must be evaluated consistently across the capability lifecycle. Existing mechanisms such as Article 36 reviews, safety cases, and governance assessments typically operate in isolation, lacking an integrated engineering-oriented decision model aligned with acquisition processes. This paper presents a lifecycle-integrated Ethical-Legal Assessment (ELA) framework designed as a structured, evidence-based assurance mechanism embedded within defence acquisition gates. The framework combines non-derogable legal hard stops with a quantitative multi-axis scoring model covering legality, ethical acceptability, safety and robustness, and accountability. A formal decision logic translates assessment results into GO, CONSTRAINED GO, or NO-GO outcomes, enabling traceable and auditable risk-based authorization. To validate feasibility, a Python-based prototype implementing structured checklists, rule-based aggregation, risk identification, and risk visualization was developed. The framework was evaluated in an urban UAS-ISR scenario, demonstrating how lifecycle evidence updates axis scores and supports controlled deployment decisions.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.


