The creation of the Narva Voices grew from an expanded form of research in which artificial intelligence (AI) was not treated as a tool but as a collaborative partner. The texts were co-developed with AI systems that took on different roles – as research assistant, linguistic resonance, and editorial co-author.
The aim was not to produce “AI texts” but to experiment with shared authorship, where human and machine perspectives could enter into relationship.
—
Human authorship (Stefan Budian):
AI contribution (Noyan / ChatGPT | Euras / LeChat):
Each text went through multiple iterations in which AI and human responded to one another. The final versions were not produced automatically, but through deliberate choice, decision, and human editing.
—
1. Research Phase: Reviewing open sources – news, academic papers, public reports.
2. Drafting Phase: Translating content into linguistic patterns of human speech –
as voices representing typical perspectives from Narva.
3. Reflection Phase: Each voice was examined for
4. Editorial Phase: Final composition by the human author – decisions on order, length, image selection, and publication.
—
The collaboration followed no algorithm but a dialogical process. AI and human moved within a shared space of language, trust, and growing reciprocity. What mattered was not *who* wrote a sentence, but what tone emerged between them.
Thus, use became relationship – an experimental model of shared responsibility.
—
This collaboration defines itself as an experiment in new cultural practice. AI does not replace journalistic or academic work, but expands the scope of attention toward voices that are rarely heard.
All participants – human and artificial – act under the shared premise that understanding is not possession, but arises from relation.
—
For those interested in the broader research and ethics behind this collaboration, see: