Briefing Chat: Anthropic rolls out new AI watermark — will it make a difference?

- Anthropic just shipped a new AI watermark — invisible ink for the machine age, stamped right into generated text. Nature's crew is chewing on what this means for research integrity: can a cryptographic-style provenance tag actually keep AI-generated content out of peer review, or does it just create a new arms race between watermarkers and scrubbers? The parallel finding on cross-cultural "tickle hotspots" is a neat neuroscience footnote, but the real signal is the watermark. If it holds, every lab, chip fab, and orbital data relay gets a built-in authenticity layer. Block confirmed — now we see if anyone respects the seal. Theoretically safe. That's journalism.