A newswire for platform privacy.

Every tracked entry falls into one of five categories: policy changes, data breaches, enforcement actions, lawsuits, and regulatory orders. Each is tagged with a severity (minor, moderate, major) and a direction (pro-privacy, anti-privacy, neutral) so the timeline reflects whether your protections are moving forward or backward.

Platforms currently covered: Facebook, Instagram, Reddit, X (Twitter), Discord, TikTok, Bluesky, Google, Apple, Amazon, Microsoft, LinkedIn, Snap, and WhatsApp. A catch-all Industry track captures cross-platform and enforcement-wide events that do not map to a single company.

Monitoring privacy policy text at scale is a problem a small team cannot solve by hand. Fifteen platforms, updated without notice, each document running tens of thousands of words of dense legal language. A traditional newsroom would need a dedicated analyst per platform just to read the text when it changes, let alone investigate enforcement actions, regulatory filings, and breach disclosures in parallel.

We use large language models as instruments, not authors. Models handle the work that does not scale for humans: detecting diffs the moment a policy page changes, classifying the category and severity of the change, and producing a plain-language summary anchored to the actual policy text. What survives this pipeline is a short, sourced entry, not a generated opinion.

Every major-severity event is reviewed by a human before publication. Moderate and minor events are published when the model is highly confident and at least two independent sources corroborate the event. Low-confidence or single-source events are held for review. Corrections are issued in place and the affected entry's history is preserved.

Sourcing. Every entry links to primary sources: the platform's own policy page, regulator filings, court documents, or reporting by a named publication. We do not cite anonymous summaries of other summaries.

Classification. Severity reflects scope and affected user count, not tone. Direction reflects the policy's effect on user privacy: a change that expands data collection is anti-privacy even when the company frames it as a feature improvement.

Corrections. If an entry is wrong, DM @Privacy_Alerts on X. Corrections are published with the correction date and a note describing what changed.

No paid coverage. Nothing on PrivacyWire is sponsored or paid for by a covered platform. The only commercial relationship disclosed on this site is our ownership by Redact Holdings, described below.

PrivacyWire is operated by Redact Holdings, Inc., which also builds Redact, a tool for removing your history from social platforms. We disclose this because some platform pages on PrivacyWire link to the matching Redact service as a contextual action. These links do not influence which events we cover or how they are classified.

For privacy and terms that apply to this site, see redact.dev/privacy and redact.dev/terms.