Making the already-disclosed visible
Google, YouTube, and some social platforms now disclose AI-generated or altered content themselves, often in small, easy-to-miss text. Noise Guard collects these existing disclosures in one visible place. Off by default.
What Noise Guard actually does
Noise Guard looks exclusively for explicit disclosures set by the platform itself, not for suspected writing patterns. Currently detected:
The disclaimer shown on AI-generated search answers.
YouTube
The self-disclosure for "altered or synthetic content".
Instagram & Facebook
The "AI Info" label on flagged posts.
TikTok
The "AI-generated" label, or "creator labeled as AI-generated".
X (Twitter)
The "Made with AI" label.
The "AI modified" label.
You choose what happens: Observe (visible only in Lens and Diagnostics, no action), Flag (a subtle notice right on the page), or Hide (currently only effective where a matching element exists, see below).
Why a disclosure is easy to miss
These disclosures already exist, usually for regulatory or transparency reasons. But they're often small, in pale grey, or shown only briefly. Anyone scrolling or reading quickly may not consciously register them at all.
Noise Guard doesn't add anything new here. It only makes visible what the platform itself already disclosed, and lets you decide whether it should stand out to you.
What this feature explicitly doesn't do
- No judgment of its own about whether text, an image, or a video looks AI-generated. Only explicit, already-set disclosures count.
- No cloud AI, no external service. The check runs with local text patterns directly in the browser.
- "Hide" deliberately doesn't hide transparency disclosures themselves, only where a matching element technically exists. Hiding a disclosure would defeat the purpose of the feature.
- No judgment call: an AI disclosure doesn't automatically mean "worse content", just more information for your own assessment.