Sarajevo research event examines AI literacy and trust in generated media

A Sarajevo research event examined AI literacy and trust in generated media.

An October 26 event in Sarajevo brought together media and regulatory stakeholders to discuss AI and information integrity in Bosnia and Herzegovina. The Council of Europe reported that the event presented preliminary results from a representative population survey on AI literacy and information integrity. Mediacentar Sarajevo also described its own research on the effect of AI-generated media content on audience trust. The projects ask related questions. The preliminary population survey and Mediacentar's audience-trust research should be described separately. [Council of Europe event account](https://www.coe.int/en/web/sarajevo/-/artificial-intelligence-media-and-information-integrity-in-bosnia-and-herzegovina) · [Mediacentar research presentation](https://www.media.ba/bs/event/prezentacija-istrazivanja-ai-i-mediji-nova-tehnologija-staro-pitanje)

Using AI and recognizing generated content are different skills

A person might regularly use an AI tool for a practical task and still have difficulty identifying how a news image or article was made. A survey can ask respondents what they believe they can recognize. A test that shows them examples and records their choices measures something else. Likewise, asking whether people trust a piece of media is different from observing whether a disclosure changes how they read or share it. These distinctions matter before applying any survey finding to a newsroom’s labeling policy.

Mediacentar describes its work as part of a broader media-literacy initiative and says the research took place in July. The Council of Europe account identifies the representative survey as preliminary. Neither source should be treated as evidence of audience behavior in other countries or under every type of AI disclosure. [Mediacentar project description](https://www.media.ba/en/project/understanding-impact-ai-generated-media-content-audience-trust) · [Council of Europe event account](https://www.coe.int/en/web/sarajevo/-/artificial-intelligence-media-and-information-integrity-in-bosnia-and-herzegovina)

What a newsroom can learn from the distinction

Editorial policies can make the provenance of a story easier to understand by explaining where AI materially affected text, images or verification. To assess whether a disclosure works, a newsroom can check whether readers understand what was generated, what journalists verified and how to report a problem. Clear language and placement matter more than a label that readers cannot interpret.

The Sarajevo event brings these questions into a specific media and regulatory context. For reporting on the studies, the next step is to read the full findings and distinguish self-reported attitudes, observed judgments and recommendations. [Mediacentar research presentation](https://www.media.ba/bs/event/prezentacija-istrazivanja-ai-i-mediji-nova-tehnologija-staro-pitanje)