OpenAI Watermarks ChatGPT Text for New EU AI Act Compliance

What happened

OpenAI is introducing invisible machine-readable watermarks for text generated by ChatGPT and Codex, specifically for users in the European Union. This move is a direct effort to align with the stringent new regulations outlined in the EU's incoming AI Act.

OpenAI rolls out invisible text watermarks for ChatGPT in the EU, aiming to comply with new AI laws, though the system isn't foolproof for detection.

Why it matters: The race to differentiate human creativity from AI-generated content just got its first major regulatory checkpoint, but perfect detection remains elusive.

• OpenAI equips ChatGPT and Codex with invisible text watermarks for EU users.
• The new 'textGrain' system aims to comply with the EU's demanding AI Act regulations.
• While designed to detect AI-generated content, OpenAI admits the watermarks are not foolproof.

Read our summary: OpenAI is adding text watermarking in ChatGPT and Codex

Open the interactive story

More Neural Snaps · Neural Toons · The Neural Feed