OpenAI explains its response to EU text provenance rules

On 5 October 2026, OpenAI explained its response to EU text provenance rules through official channels. According to the report, OpenAI's text watermarking approach applies to some scenarios, the detection mechanism is already defined, and access is first opened to researchers. That is all the currently confirmable fact: the event centres on OpenAI's compliance response to EU text provenance rules, with text watermarking as its core measure, but the report does not disclose the specific range of scenarios the watermark covers, the technical details of the detection mechanism, or when parties other than researchers will gain access. Whether there were earlier related reports, or any inconsistency in figures or statements, is not provided by the available material, so no comparison is possible. Further developments await more disclosure.

Written by AI from the reports · updated 11 d ago