Ethical Guidelines for the Development, Application and Use of Reliable and Responsible Artificial Intelligence

Summary

Non-binding guidelines adopted by Government Conclusion on 23 March 2023 as a preventive mechanism for checking that machine-learning systems meet the highest ethical and security standards. The Government recommends that state, provincial and local bodies, public institutions and enterprises, regulators and holders of public powers apply them to the AI systems they set up and use, and other developers and users too. It defines AI systems and high-risk AI systems, lists high-risk areas such as biometrics, critical infrastructure and law enforcement, and sets four principles and seven conditions for reliable and responsible AI, with recommendations and an assessment questionnaire.

Language

No English version of this instrument was found. The English title and the summary are the register's own, written from the sources below; for the exact wording use the Serbian text.

Safety topics

Status notes

The four principles are explainability and verifiability, dignity, prohibition of harm and fairness. The seven conditions are agency and oversight; technical reliability and safety; privacy, personal data protection and data governance; transparency; diversity, non-discrimination and equality; social and environmental well-being; and accountability. The guidelines rest on the AI Development Strategy for 2020-2025 (Službeni glasnik RS 96/2019), one of whose five goals is ethical and safe AI, and its Action Plan for 2020-2022 (Službeni glasnik RS 81/2020), and draw on the 2021 UNESCO Recommendation on the Ethics of AI and the European Commission's April 2021 AI Act proposal. The competent minister determines how their application is integrated. They do not cover systems prohibited by a law regulating AI, and they record that AI needs to be regulated by law. The text is dated February 2023.

Sources

Open the official source https://www.srbija.gov.rs/vest/en/203799/ethical-guidelines-for-development-use-of-artificial-intelligence-adopted.php

Secondary source: https://kmp.respaweb.eu/download/doc/Ethical%20Guidelines%20for%20Development%20Application%20and%20Use%20of%20Artificial%20Inteligence.pdf/e841c4ca3e9354bf0e6191eea34d1ee6.pdf

Local copy

Verification

History

1. 2026-10-17english_version: – → {"kind": "none found"} — no English translation of the text was found; the English title comes from the Government's own English news item2026-10-17-recheck-europe-3 · Claude Code (claude-opus-5-5)
2. 2026-10-17notes: These remain Serbia's only adopted AI-specific governance instrument with safety → The four principles are explainability and verifiability, dignity, prohibition o — rewritten from the text: principles, conditions, basis, minister's role, prohibited systems excluded, need for a law; process remarks moved to verification.note and unstored claims removed2026-10-17-recheck-europe-3 · Claude Code (claude-opus-5-5)
3. 2026-10-17safety_topics: ["risk-based classification", "transparency & labelling", "accountability & liab → ["risk-based classification", "human oversight", "transparency & labelling", "ac — 'human oversight' (condition 1, agency and oversight) and 'public-sector use' (recommended to all public bodies) added2026-10-17-recheck-europe-3 · Claude Code (claude-opus-5-5)
4. 2026-10-17summary: Non-binding guidelines adopted by Government Conclusion on 23 March 2023 as a pr → Non-binding guidelines adopted by Government Conclusion on 23 March 2023 as a pr — rewritten from the Serbian text: the Government recommends the guidelines to public bodies (1.2), it does not make them apply; the four principles, seven conditions, high-risk areas and the assessment questionnaire added2026-10-17-recheck-europe-3 · Claude Code (claude-opus-5-5)

Related records

UNESCO AI Ethics Recommendation UNESCO · Published (non-binding, current)

Serbia AI Strategy 2025-2030 Serbia · Published (non-binding, current)

Citation

Government of the Republic of Serbia (adopted by Government Conclusion). Ethical Guidelines for the Development, Application and Use of Reliable and Responsible Artificial Intelligence. 2023-03-23. https://www.srbija.gov.rs/vest/en/203799/ethical-guidelines-for-development-use-of-artificial-intelligence-adopted.php. Status: Published (non-binding, current); verified 2026-10-17 (level A). Local copy archive/RS-01/2026-10-17/official.html retrieved 2026-10-17 20:47:58 UTC, SHA-256 0cc9786d2cc463ce385abcaf269b100b8aabfbd0e44f4261c1d83b060188e519. Every AI Policy Ever, version 1.18 (snapshot 2026-10-18), record RS-01.

Improve this record

Know of an amendment, a status change or an error? Report news or an error on GitHub; the form opens with this record filled in.