RESEARCHInvestigateNEXT 12 MONTHS
Proxy reliance in large language model decisions is uncalibrated to predictive evidence
arXiv cs.CL — Computation and Language
Factual evidence
What the source reports
Research shows LLM decision-making conflates protected proxy attributes with predictive evidence, undermining standard demographic bias audits.
Inspect the evidence
- Inclusion basis
- Enterprise AI
- Publisher and source type
- arXiv cs.CL — Computation and Language · RESEARCH
- Published by source
- 22 October 2026
- Collected by OneBench
- 26 Aug 2026, 03:04 UK
Stored source excerpt
arXiv:2608.22887v1 Announce Type: cross Abstract: Large language models (LLMs) are entering decisions in triage and lending, where task-relevant inference must be distinguished from impermissible proxy…
Short excerpt from the collected text, not the full source. Use the source link to read it in context.
The factual summary is a OneBench synthesis, not a quotation or independent verification. Collection time is not publication time. Open the source for its full context; related reporting can share the same underlying announcement.