Editorial routing shapes how computational results are qualified in AI-assisted scientific writing

arXiv:2609.14288v1 Announce Type: new Abstract: Large language models increasingly analyze computational results and draft manuscripts, making reliable communication as important as correct analysis. Using fixed computational evidence, we tested whether assigning comparisons across modeling choices elsewhere in a research workflow changes manuscript reporting. In constrained sentence-writing tasks, Anthropic's Claude Sonnet 5 often omitted numerical qualifications when detailed comparisons were assigned to a group repository, but retained them more often when the same comparison was assigned to Supporting Information or its own working notes; Claude Opus 5 was less sensitive. These effects did not follow a simple accessibility ordering. A targeted placement rule largely restored sentence-level qualification, whereas a generic accuracy reminder did not. Longer contributions retained numerical qualifications, although some summaries across computational settings were still redirected to the repository. Thus, documenting context within an AI workflow does not ensure its communication where readers encounter the result.

Related

• Source or It Didn't Happen: A Multi-Agent Framework for Citation Hallucination Detection
• Evaluating and Calibrating LLM Confidence on Questions with Multiple Correct Answers
• Claim-Level Confidence Calibration for Reliable Decision Making with Large Language Models
• PeerPrism: Peer Evaluation Expertise vs Review-writing AI
• MeasHalu: Mitigation of Scientific Measurement Hallucinations for Large Language Models with Enhanced Reasoning

Source: arXiv cs.CL | 2026-09-15