How This Briefing Works
This report opens with key findings, then maps the gaps between what Google Gemini discloses and what BLACKOUT observed at runtime. From there: what it means for your organization, what to do about it, and the detection data and evidence underneath.
Key Findings
Pre-Consent Activity
Google Gemini was observed loading and executing before user consent was obtained on 6% of sites where it was detected.
Claims vs. Observed Behavior
pending
“Requires claims extraction via CDT”
Live website analysis pending
What This Means For You
What To Do About It
Role-specific actions based on observed behavior
If You're Evaluating Google Gemini
- →Self-hosted LLM alternatives
- →Anthropic Claude API (no training on customer data)
- →OpenAI API with zero-retention configuration
Negotiation Leverage
- →Google cannot contractually limit data usage — their business model requires cross-platform profiling and advertising integration
- →Consent bypass mechanisms (C01, C09) create strict liability regardless of DPA terms
- →Behavioral biometrics and session recording trigger GDPR Article 9 special category requirements that standard agreements do not address
- →Demand signal leakage (100% broker score) subsidizes competitors using Google advertising products
Runtime Detections
BLACKOUT observed this vendor's JavaScript executing in a live browser and classified each hostile behavior using our BTI-C (Behavioral Threat Intelligence — Capability) taxonomy. These are not theoretical risks — each code below was triggered by something we watched this vendor's code actually do.
Evasion infrastructure, auditor bypass
Impact: Active evasion of consent mechanisms ensures tracking proceeds regardless of user choices, creating strict liability under consent frameworks.
Keystroke/mouse tracking
Impact: Interaction pattern capture with AI-powered analysis enables persistent re-identification even after identifier deletion.
Full session replay
Impact: User interactions with Gemini interfaces captured and transmitted to Google for training data and behavioral profiling.
Identity stitching
Impact: Gemini user IDs synchronized across Google properties, enabling cross-site tracking and profile enrichment.
Ignoring CMP signals
Impact: Data collection initiates before consent frameworks load, capturing pre-consent user behavior.
Device identification
Impact: Browser and device characteristics harvested to create persistent identifiers independent of cookie consent.
Long-lived identifiers
Impact: Multiple storage mechanisms (localStorage, indexedDB, cache) ensure identifier survival across browser sessions and clearing attempts.
PII deanonymization
Impact: Gemini user activity linked to Google advertising profiles, enabling deterministic cross-platform tracking.
Container/loader (neutral)
Impact: Google Tag Manager used to load Gemini tracking outside consent scope, evading blocking attempts.
IOC Manifest
Indicators of compromise across 4 categories. Use for detection rules, CSP policies, or Pi-hole blocklists.
Ecosystem & Supply Chain
Evidence Artifacts
Artifacts collected during analysis, available with evidence-tier access.
Complete network capture with all requests and responses
42 detection signatures across scripts, domains, cookies, and network endpoints