OpenAI is adding text watermarking to ChatGPT and Codex, starting in the EU to comply with the EU AI Act and offering API opt-in globally.
6 stories from 6 sources · since Oct 5
OpenAI is adding invisible text watermarking to ChatGPT and Codex, primarily to align with the EU AI Act, with optional API support offered globally. The move pushes AI content provenance and compliance from policy into product features. CIOs should care because it changes governance, auditability, and regional control requirements for generated text, while the watermarking itself remains imperfect and bypassable.