Gemini suddenly refuses to extract text from scanned books, citing copyright

Mr_VRBeerscuit · reddit · 2026-09-22

A user reports Gemini was extracting text from a scanned book fine until a few hours ago, when it suddenly began refusing with a "this is copyrighted content" warning. The post adds to recent complaints about overly tight guardrails blocking legitimate OCR-style use cases.

Original post →

More from Models

Models channel →