Portugal

ChatGPT will identify texts with invisible watermarks in the EU

Read full article at source ↗

Based on a feed excerpt; full article content has not been acquired. Source evidence retrieved 06/10/2026, 12:05:25 UTC.

Briefings

The OpenAI system introduced in October 2026 embeds invisible watermarks into AI-generated text to comply with the EU AI Regulation, allowing machines to detect these markers while humans cannot. This marks a shift from purely human-readable outputs to machine-detectable digital signatures for regulatory compliance.

Key points

• The OpenAI system embeds invisible watermarks into AI-generated text to comply with the EU AI Regulation.
• The system allows machines to detect these markers while humans cannot perceive them.

Context

This finding addresses the EU AI Regulation's requirement for machine-readable compliance markers in AI-generated content, enabling automated verification of AI-generated text.

Limitations

Evidence is limited to a single article from October 2026; no data on implementation scale, detection accuracy, or specific text examples is provided.

Generated by Bittrees-hosted qwen35-2b sha256:aaf42c8b7c3c on 20/10/2026, 10:22:51 UTC. Based on publisher-provided feed evidence; may not cover the full original article. Check the source for details.

View archived briefing on IPFS
bafybeiakou26i5cfr7eysderqgwrz2piwacr4nxchyu2fkdf6jpnphonmq