VARNIUMAI risk defense / everyday bearings

PRIVACY / OPT-OUT

Opt Your Voice and Photos Out of AI Training, Step by Step.

Your voice and likeness are the raw material of AI generation, and several major platforms have opt-out routes. The finder turns the question "can I opt out" into a dated checklist.

Start here

The free tool below runs entirely in this browser — no account, nothing stored or transmitted. Redact anything you would not want kept before you type it.

Open the companion tool →

Why partial still matters

There is no single switch that removes your voice and likeness from every training set. But most major platforms offer some route — an opt-out toggle, a data request, a removal form — and each one reduces the raw material available to generators.

The honest framing is reduction, not erasure: each completed opt-out narrows where your likeness can be scraped, and the dated record lets you re-check on a schedule instead of assuming it stayed done.

The protocol

  1. Select the platforms where your voice or photos are hosted.
  2. Follow the per-platform route — toggle, request, or form — and record the date.
  3. Keep the confirmation (screenshot or email) with the date.
  4. Set a follow-up date to re-check, since routes and settings change.

What the tool does

Select the platforms holding your voice or likeness and get a dated opt-out checklist with follow-ups. It runs in this browser; nothing is sent or stored.

Limit first: This is a checklist aid, not a guarantee that any opt-out will be honored, and not legal advice. Routes change; verify each platform's current form.

Open the companion tool →

One human next step

Complete the first platform's opt-out today and keep the confirmation. Re-check routes periodically, and file a privacy complaint with the FTC if a platform ignores a request.

Research log and safety checks

AI-assist path: not applicable — opt-out routes are platform-specific settings and forms the reader must complete themselves; the artifact is the checklist and the record, not a model judgment. §6.10: not a frontier-pace topic.

§0.5 protection result: category is data-privacy-exposure; crisis-adjacent: no; fear/urgency/scarcity toolkit used: no. The protective function is free and the paid feature is convenience only.

Related dispatches

Pairs with the same protection bar: a free complete reading, an optional convenience unlock, and no fear-framing.