Text Provenance
A set of technologies and standards for tracing and verifying the origin of textual content, used to determine whether content was generated by a human or an AI system. It forms an important component of the European Union's AI regulatory framework.
Timeline (last 90 days)
Text provenance rules recently advanced by the EU require traceable source identification for AI-generated or AI-assisted text.
Tracing provenance for plain text is harder than for images or audio, because existing markers are easily lost once text is copied, rewritten, or translated.
The three main technical approaches to text provenance are metadata annotation, statistical watermarking, and explicit declaration.
The EU has introduced text provenance rules requiring AI service providers to offer traceable and identifiable technical means for content generated by models.
All Facts (4)
Text provenance rules recently advanced by the EU require traceable source identification for AI-generated or AI-assisted text.
50%UnverifiedTracing provenance for plain text is harder than for images or audio, because existing markers are easily lost once text is copied, rewritten, or translated.
50%UnverifiedThe three main technical approaches to text provenance are metadata annotation, statistical watermarking, and explicit declaration.
50%UnverifiedThe EU has introduced text provenance rules requiring AI service providers to offer traceable and identifiable technical means for content generated by models.
50%