Study: LLM pushback is so ambiguous that users still believe false content after being corrected
lucy3_li · x · 2026-10-14
A UW/Stanford/AI2 paper, 'Pushback in a Haystack', shows LLMs that do push back against false user inputs do so ambiguously — indirect phrasing mixed with affirmation, strongest in Hindi and Yoruba. A 300-person human study finds this ambiguous pushback often fails to correct users' factual misconceptions: sycophancy isn't solved just because models object.
Related event: AI Models Push Back Weakly on False Inputs, Users Still Misled(2 posts)→
More from Research
- RFD4-Proteina's 'underrated' docs detail partial diffusion for protein redesign — AllThingsApx · 2026-10-23
- MPI-Sec PhD Thesis on Vulnerability Detection Benchmarks Includes Pre-LLM Classic "Top Score on the Wrong Exam" — mboehme_ · 2026-10-23
- LiveNerf's 29-day data finds no nerfing of Claude Opus 5.5 — TheOnlyVibemaster · 2026-10-23
- New paper makes 9 falsifiable predictions toward a major AI-enabled terror attack by 2027-2028 — StephenLCasper · 2026-10-23
- HumanoidPose recovers humanoid robots' full-body joint states from RGB video alone — cholz · 2026-10-23
- Sophont's Holon-S1 predicts genes and biomarkers from routine pathology slides — iScienceLuvr · 2026-10-23