← The Polylog AI Intelligence Brief
Morning Edition · Wednesday, October 7, 2026Published at 2:04 AM EDT · New York
OpenAI will watermark ChatGPT and Codex text in the European Union to comply with the AI Act
The statistical watermark, called textGrain, was detected in about 95% of 400-token passages in OpenAI's tests, but replacing one word in ten with a synonym cut detection to about 66%.
OpenAI will add an invisible watermark to text that ChatGPT and Codex generate for users in the European Union, TechCrunch reported. The company expects the change to reach eligible users on all EU plans over the coming weeks. The trigger i…
Continue the AI Intelligence Brief
Track frontier labs, chips, export controls, model releases, regulation, and AI infrastructure.
- 5 AI intelligence signals a day
- Frontier labs, compute, and chips
- Model releases and AI infrastructure
- Source-grounded analysis with confidence labels
The Global Intelligence Brief stays free.
Part of a tracked trend
State-Mandated Provenance Marking of AI Output
Regulators will keep requiring machine-detectable marks on AI output, which concentrates detection power with model providers and expands the traceability of private communications more than it proves where any given text came from.
Townsquare
Start a discussion in Townsquare.
More from this edition
- Microsoft and Meta cut internal Claude spending weeks before Anthropic's expected listing
- OpenAI publishes 722 math manuscripts from an unreleased model, with Lean proofs for part of the set
- Wikimedia links OpenAI agents to unauthorized wiki edits and attempts to use its tools as proxies
- No frontier lab tells New York City Council it carries catastrophic-risk insurance
- Google releases EmbeddingGemma 2, a 740 million parameter open model that embeds text, code, images, video and audio
- AWS and Ai2 ship small open specialist models built on Alibaba's Qwen base weights
- Three new papers move prompt-injection defense for agents to the point of action
Comments
0No comments yet.