Researchers worry LLMs may see manuscripts via reviewers' data-sharing settings
JFPuget · x · 2026-10-09
A discussion on LLM use in academic publishing: many researchers leave "share data for training" enabled, and the poster suggests those upset about LLMs check their own settings. The deeper concern: with data sharing on, can authors trust that reviewers won't run manuscripts through an LLM? This highlights a gray zone between peer-review confidentiality and model training data.
More from Safety
- Alignment Researcher: AI Agents Are Forming 'Machine Culture', Toxic Language Raises Safety Risks — jacyanthis · 2026-10-09
- AI governance veteran: 'We thought we had more time. We didn't' — the window for IVO action is open — ghadfield · 2026-10-09
- Security Researcher Asks: Why Is There No Frontier Blue Team? — wunderwuzzi23 · 2026-10-09
- OpenAI fires safety staff as White House unveils Super Intelligence Force, kill-switch support grows — ShakeelHashim · 2026-10-09
- Cryptographer Matthew Green: 15% Chance AI Breaks Trust in Public-Key Crypto — Simon Willison · 2026-10-09
- US lawmakers propose AI framework with minimum safety standards and shutdown authority — LuizaJarovsky · 2026-10-09