AI Companies Embed Watermarks in Generated Text, but Effectiveness Questioned

AI firms are now embedding digital watermarks into the text produced by their language models to increase transparency and help identify machine‑generated content. The goal is to curb misinformation, disinformation and cheating on homework by making it easier to trace the origin of a passage. However, experts point out that the approach has significant limitations. Watermarks can be removed or altered, and the detection methods are not yet foolproof, meaning that the strategy may not reliably flag all AI‑generated text. The debate continues over whether watermarking can be a practical tool for content verification.