← All pages

1. What the document is: Google's unified Privacy Policy plus the Gemini Apps Privacy Notice, covering data collection, AI-training data use, retention, and a detailed Gemini-specific FAQ (chats, photos/videos, audio/Gemini Live, screen automation, Connected Apps, imported chats from other AI platforms).

What this wiki found — complete, every page

Presence facts, not verdicts — each flagged term links to its definition and the exact clause on this page. Absence of a badge means the term isn't currently flagged here, not that the page is risk-free.

Independent reviews, where they had something to say

Google Generative AI Services – Privacy Policy

Dataset: GenGA (Generative AI Governance Archive) — 11 AI services, 2025–present

GenGA (Generative AI Governance Archive) is this wiki’s Generative-AI dataset: raw policy snapshots captured by the Open Terms Archive genai-eu project across 11 GenAI providers (ChatGPT, Claude.ai, DeepSeek, Google Generative AI Services, Le Chat, Llama API, Meta AI, Microsoft Copilot, Perplexity, Qwen Chat, xAI). Unlike PGAv2, GenGA has no pre-tagged risk scores — all risk scoring and keyword tagging on these pages is LLM-assigned by direct reading, spanning 2025–present.

Note: This page contains documented policy clauses. Risk assessment is qualitative and context-dependent. For analysis of patterns across platforms, see:

• AI Act Substance Matrix
• Tier Bifurcation
• Scraper Enforcement Gap
• Methodology

Source: sources/GenGA/Google Generative AI Services/Privacy Policy/ (raw Markdown captures, Open Terms Archive genai-eu dataset — no pre-tagged JSONL exists for this source) Last updated: 2026-04-17

Note on methodology: No pre-computed risk_score/keyword fields exist for this source. Keyword tags and clause analysis on this page are LLM-assigned, using the same rubric/vocabulary as the PGAv2 pages plus new GenAI-specific tags (input license, gdpr). Treat as first-pass analysis, not externally verified ground truth. This wiki’s risk-scoring system was retired project-wide on 2026-06-21 (see methodology.md §4) — no page, GenGA included, computes or displays a numeric risk score.

Document-bundling note: Like several other captures in this dataset, this file bundles two documents: Google’s main unified Privacy Policy (GDPR-style legal-basis tables, the same document covering all Google services) and the Gemini Apps Privacy Notice (Gemini-specific data-use FAQ). Findings below are drawn from both.

Overview

1. What the document is: Google’s unified Privacy Policy plus the Gemini Apps Privacy Notice, covering data collection, AI-training data use, retention, and a detailed Gemini-specific FAQ (chats, photos/videos, audio/Gemini Live, screen automation, Connected Apps, imported chats from other AI platforms).

2. Input rights: Default training-eligible: “Google uses your activity to provide, develop, and improve its services (including training generative AI models)… with the help of human reviewers,” gated by a “Keep Activity” setting — but even when Keep Activity is off, Google still uses your chats to respond to you and help protect Google, our users, and the public, including with help from human reviewers. This opt-out is narrower than OpenAI’s or Anthropic’s: turning off Keep Activity stops chats from training the models, but human-reviewer-assisted “protection” use of chats continues regardless of the toggle.

3. Output restrictions: Not addressed in this document.

4. Non-explicit predatory clauses: The most distinctive finding in this dataset: Gemini lets users import memory/chats from other AI platforms (e.g., competitor chat exports via a .zip file), and “Data (memory and chats) that you import from other AI platforms… is used consistently with your Gemini Apps activity, including to improve our services (including training generative AI models).” No other provider in this dataset has a documented “import-and-train-on-competitor-data” feature.

5. Regulatory references: GDPR explicitly named in full: “Under certain privacy laws, including the General Data Protection Regulation in the EU, you may have the right to…” The legal-basis table uses GDPR-style categories (legitimate interests, vital interests, legal obligation, consent). No EU AI Act reference found.

6. Regional variation: Significant and unusual direction — rather than adding EU-specific protections, one Gemini feature is explicitly withheld from EU/UK/Swiss users: “Outside of the European Economic Area, Switzerland, and the UK, Gemini personalizes your experience by referencing your chats in Gemini Apps Activity” (the “Memory”/personalization feature) — implying EEA/Switzerland/UK users do not get this feature at all, rather than getting a modified version of it.

7. Key risk to users: The “always-on” human-reviewer-assisted chat processing that survives the Keep Activity opt-out is the central risk, compounded by the cross-platform data-import-to-training pipeline and the audio/Gemini Live/screen-automation screenshot collection (also human-reviewed, also tied to the same toggle).

Flagged Keywords & Risks (LLM-assigned)

• input license, train AI/models — chats, audio/Gemini Live recordings, screen-automation screenshots, and chats imported from competitor platforms are all used “to improve our services (including training generative AI models).” This is gated by the Keep Activity setting, but a carve-out lets Google keep using chats “with help from human reviewers” even when Keep Activity is off. Why it matters: the narrowest effective opt-out found in this dataset — a toggle that doesn’t fully stop data use even when switched off.
• gdpr — explicit “General Data Protection Regulation” citation plus a GDPR-style legal-basis table. Why it matters: concrete regulatory grounding for the cross-provider comparison.

Regulatory & Research Context

Pandit et al. (2026), who directly studied Google/Gemini’s terms, found that “all terms mentioned that inputs and outputs would also be used for other purposes beyond” the immediate service — a finding this page corroborates directly, since even with Keep Activity switched off, Google “still uses your chats… with help from human reviewers.” Davidson et al. (2026), also direct Google researchers, frame such narrow, only-partially-effective opt-outs as part of the “regulatory gray areas” that complicate legitimate use and consent; the cross-platform “import chats from other AI platforms… including to improve our services (including training generative AI models)” feature flagged on this page extends that gray area into competitor-data ingestion, a pattern not documented for any other provider in this dataset.

Changes Summary

Version History

2026-02-20

• Explicit AI clause: YES
• Non-explicit predatory: NO
• Flagged keywords: input license, train AI/models, gdpr

Clause: input license, train AI/models

Google uses your activity to provide, develop, and improve its services (including training generative AI models), as well as to protect Google, its users, and the public with the help of human reviewers. Data reviewed by service providers is disconnected from your account and saved for 3 years. Your audio and Gemini Live recordings aren’t used to improve Google services by default. Depending on your settings and region, Google also uses your activity to personalize your experience.

Clause: gdpr

Under certain privacy laws, including the General Data Protection Regulation in the EU, you may have the right to: […]

2026-03-10

gdpr unchanged from 2026-02-20.

Clause: train AI/models (new — screen automation)

During automation, Gemini takes screenshots, which can include any info visible in the app. Screenshots are reviewed by trained reviewers and used to improve Google services if Keep Activity is on. Don’t enter login or payment information into Gemini chats.

2026-03-20

train AI/models/gdpr unchanged from 2026-03-10.

Clause: regional variation (renamed, not separately concept-tagged)

Outside of the European Economic Area, Switzerland, and the UK, Gemini personalizes your experience by referencing your chats in Gemini Apps Activity and using them to generate personalized insights. You can turn off Memory anytime to stop personalization based on your past Gemini chats. Keep Activity must be on to use this feature.

2026-03-26

train AI/models/gdpr unchanged from 2026-03-10.

Clause: input license, train AI/models (new — cross-platform data import)

Data (memory and chats) that you import from other AI platforms is saved in your Activity. This data is used consistently with your Gemini Apps activity, including to improve our services (including training generative AI models), and to protect Google, our users and the public. You can manage or delete your activity anytime.

2026-03-27

All flagged clauses unchanged from 2026-03-26 — this date’s changes were minor wording additions referencing “imported chats” in adjacent sentences.

2026-04-02

All flagged clauses unchanged from 2026-03-26 — this date’s changes were to the bundled main Privacy Policy’s effective date and unrelated cookie/Incognito-mode wording.

2026-04-08

All flagged clauses unchanged from 2026-03-26 — this date added a NotebookLM cross-reference, not a flagged clause.

2026-04-09

All flagged clauses unchanged from 2026-03-26 — this date’s only change was a date-stamp synchronization.

2026-04-10

All flagged clauses unchanged from 2026-03-26 — this date’s only change was a date-stamp synchronization.

2026-04-16

All flagged clauses unchanged from 2026-03-26 — this date added an Android-app-management help link, not a flagged clause.

2026-04-17

All flagged clauses unchanged from 2026-03-26 — this date added a new “Connected Apps” personalization FAQ entry, not itself re-quoted as it describes an opt-in personalization feature rather than a new training/license clause.