IT之家 · ITHome
SpaceXAI to add invisible watermarks to Grok text, citing EU AI Act
SpaceXAI says it will add invisible watermarks to Grok-generated text from November 7 in response to the EU AI Act, and that its APIs worldwide already have watermarking enabled. The signal is embedded through token selection rather than hidden characters or metadata. Detectors can identify the signal, while team administrators can disable watermarking for individual models.

What to know
- 01
Detectors identify the watermark signal to determine whether text carries the invisible marker.
- 02
Team administrators can turn off watermarking for an individual model.
Why it matters
The update concerns watermark detection in Grok text and SpaceXAI’s stated response to EU AI Act requirements.
AI-assisted brief. Read the source for full details. Editorial & sourcing policy


