Sleuth AI is an AI research integrity assistant built into Signals Manuscript Checks, which lets editorial and integrity teams ask questions of a manuscript and surface the evidence. In its first release, which uses pre-defined prompts, it extracts and highlights conflict-of-interest, funding and ethics statements and flags anything unusual, shows where potential citation misconduct occurs, and runs exploratory checks for irrelevant or inconsistent text that may indicate low-quality or AI-generated content. Signals announced it on 21 July 2025 as available to publishers using Signals Premium, through the ScholarOne integration and direct upload.
Notes & quotes
- Signals says it designed Sleuth AI to say when it lacks an answer.
- Signals plans to connect Sleuth AI to the Signals Data Graph, for example to explain how a retracted reference affects an article’s conclusions.
Publisher / company: Signals
Date of case study: 2025
Status: Live
Built by: Company
Sources: