Summary

The United Nations Industrial Development Organization (UNIDO) presented recommendations for the adoption of artificial intelligence (AI) in industry. As AI is increasingly used to manage production processes and logistics, technical failures may affect not only individual enterprises but also entire supply chains. Experts therefore propose regulating not individual algorithms, but the entire system in which they operate — from data and equipment to staff actions and the allocation of responsibilities. The approach is based on six interrelated principles: safety, reliability, transparency, accountability, inclusiveness and sustainability. The model covers the international, national and corporate levels. At the global level, it provides for the harmonization of standards, information-sharing on incidents and mutual recognition of conformity-assessment results. Governments are advised to maintain digital risk registers and integrate such risks into existing industrial safety mechanisms. Enterprises should appoint officials responsible for the operation of AI systems, ensure continuous monitoring and address risks in their relationships with suppliers. These measures should be supported by special regulatory regimes, digital testing platforms and independent audits. Implementation of the approach is designed to take place over five years, beginning with the establishment of national registers and incident-reporting mechanisms. The next stage involves developing conformity-assessment systems and sector-specific regulation, followed by expanded international cooperation. UNIDO also proposes a nine-level AI readiness scale for assessing the reliability of individual systems. At the same time, the organization emphasizes that the scale is not yet an international standard and requires further validation and adaptation to specific industries.