Does a Tool Result Carry More Authority Than Plain Text? Three Prospective Studies of False-Claim Adoption in a Synthetic Assignment Task with Claude Opus 5

Researchers conducted three studies to determine if providing unsupported claims through a retrieval-style interface influences AI decision-making compared to plain text presentation. The findings suggest that the format of information delivery can affect how models adopt false claims, highlighting potential risks as language models increasingly rely on data stores they simultaneously populate. This research underscores the importance of verifying information integrity in systems where AI-generated content becomes part of the knowledge base that future models query.