SpaceXSI will add a watermark to all text outputs from its Grok API models by default starting Nov. 7, a move the company is making to comply with the European Union's Artificial Intelligence Act.
Nov. 7 · Grok API
Grok API text outputs will carry a watermark by default
The change covers every text response from Grok API models and is aimed at EU Artificial Intelligence Act transparency rules. Developers and businesses will see a new default once it takes effect.
Nov. 7
Start date for the default watermark
All text
Every text response from Grok API models
EU AI Act
Regulatory driver behind the change
What happens next
Step 1
Default watermark applied to text outputs
→
Step 2
Developers adapt products and communications
→
Step 3
Method, detection and opt-out details published
Why it matters
Watermarking is a direct technical response to EU transparency pressure.
Generated text can carry a signal that it came from an AI model.
Providers serving European users must show their systems meet standards.
Known limits
Text is easier to edit, paraphrase or translate, which can weaken a watermark.
A stronger mark may affect output quality.
A default watermark alone is unlikely to settle questions about AI-generated text.
Still unclear
The reporting does not yet include the watermark's technical details.
Embedding
How the mark is built into text
Detection
Public or only through specialized tools
Opt-out
Whether it can be disabled in any case
Continue reading The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in ✓ Signed in — this article isn’t included in your current plan.Unlocking the full article...