VLMs Tend to ‘Correct’ Faulty Text, Threatening OCR Replacement Plans
A new study shows vision‑language models often rewrite noisy document text into a plausible version rather than copy it verbatim. Using the multilingual FaithC4 benchmark, the authors find general‑purpose VLMs’ error rates jump up to 4.5 WER points, far higher than OCR‑specialized VLMs or traditional OCR. This means swapping OCR pipelines for VLMs can silently corrupt data in real‑world workflows.