Serbian Journalists' Association Adopts AI Use Policy

The Independent Association of Journalists of Serbia (NUNS) has established a policy for the ethical and transparent use of artificial intelligence within its operations. Adopted on June 2, 2026, the policy aims to guide the responsible application of AI technologies.

“

NUNS has adopted a document whose basic principle is that artificial intelligence can improve the organization's work, but it cannot replace human judgment, journalistic ethics, and editorial responsibility. The use of these tools must be strictly controlled and based on constant human supervision.

NUNS recognizes AI's capacity to accelerate tasks like research, data processing, transcription, and translation in newsrooms. However, the association also stresses the importance of addressing critical issues such as accuracy, copyright infringement, and the protection of personal and confidential data. The core principle of the adopted document is that while AI can improve organizational work, it cannot substitute for human judgment, journalistic ethics, and editorial responsibility. The use of AI tools must be strictly controlled and subject to continuous human supervision.

The policy outlines key principles including accountability, transparency, privacy protection, informed consent, accuracy verification, and ongoing education. Individuals using AI tools remain responsible for their outcomes, and any content partially or fully generated by AI must be clearly marked. NUNS prohibits inputting personal or confidential data into AI tools and mandates independent fact-checking, source verification, and careful contextual assessment.

“

The person using an artificial intelligence tool remains responsible for its result, while any content that is partially or entirely created using such a tool must be clearly marked.

Furthermore, the policy introduces a requirement to assess the security of any new AI tool before its implementation. Decisions on using specific tools will only be made after confirming their safety, with periodic reviews of their performance, advantages, and disadvantages. NUNS notes that clear rules are particularly crucial given the low public trust in media, citing a 2026 Reuters Institute Digital News Report indicating only 22% of Serbian respondents trust the news. Serbia currently lacks comprehensive legislation specifically regulating AI use in media.

“

According to the Digital News Report 2026 by the Reuters Institute for the Study of Journalism, only 22 percent of respondents in Serbia say they generally trust the news.