ETHICAL DILEMMAS AND RISKS IN THE APPLICATION OF ARTIFICIAL INTELLIGENCE FOR INDUSTRIAL SAFETY: A SYSTEMATIC REVIEW.

Abstract

The incorporation of artificial intelligence (AI) in industrial safety in the manufacturing sector represents an emerging field with high potential for the proactive prevention of occupational risks, but it also entails ethical dilemmas and significant socio-technical risks. Through a systematic review of the literature (2015–2025) and a bibliometric analysis of 467 articles, this study identifies the main challenges related to worker privacy, fairness and algorithmic biases, autonomy and human control, as well as transparency and accountability. In addition, technical risks such as classification errors, model degradation (drift), cybersecurity vulnerabilities and reliance on edge/cloud architectures are documented. The findings reveal that Explainable AI (XAI), worker-centered socio-technical design, and the implementation of robust regulatory frameworks (AI Act, PLD, AILD) are key mitigation strategies. The study concludes that the integration of AI into industrial safety can optimize accident prevention as long as a comprehensive approach is implemented that articulates technical, organizational, and legal measures to ensure a safe, ethical, and equitablework environment.