VendorScore · AI assistant · Data-access posture
A plain-language answer from Google's Generative AI Privacy Hub and its data-regions, CSE, DPA and subprocessor pages. Only client-side-encrypted content is out of Gemini's reach.
Get the full 5-dimension PDF — join the waitlistEvidence dated 2026-10-07, public docs only, not legal advice.
Short answer
Yes, unless you enable CSE: Google's systems process Gemini prompts and the Workspace content they retrieve in plaintext; client-side-encrypted content is technically out of Gemini's and Google's reach.
Scored configuration: Gemini in Workspace apps and the Gemini app on Workspace Enterprise Plus/Frontline Plus with data regions (in-region processing and storage), DLP/IRM, and Client-side encryption (CSE) used to keep sensitive content out of Gemini.
Default setup without CSE or data regions scores lower: Google-managed keys, and EU residency isn't guaranteed. Scored separately; the Google Workspace score isn't reused.
Report confidence: Medium-high.
Higher = more customer control / less inherent vendor read access for the scored configuration. A strong SOC 2 doesn't mean the vendor can't read your data.
Workspace doesn't use customer data (including prompts) to train models without prior permission; content isn't human-reviewed for training outside your domain.
Evidence: Generative AI Privacy HubFor CSE-protected content only: Google says CSE can restrict Gemini's access because no Google system or Google employee has the technical means to access CSE content. Other content Gemini retrieves is processed in plaintext.
Evidence: Generative AI Privacy HubNo third-party AI model providers; third-party Workspace subprocessors are support vendors that see data only if you grant access in a support case.
Evidence: Workspace subprocessorsOur report flags these gaps in Gemini for Google Workspace's public documentation:
Get the full 5-dimension PDF — join the waitlist
PDFs aren't for sale yet. Join the waitlist and we'll email yours when it's ready. No calls.
Yes, unless you enable CSE: Google's systems process Gemini prompts and the Workspace content they retrieve in plaintext; client-side-encrypted content is technically out of Gemini's and Google's reach. VendorScore rates Gemini for Google Workspace 69/100 overall from public documentation (evidence dated 2026-10-07). Scored configuration: Gemini in Workspace apps and the Gemini app on Workspace Enterprise Plus/Frontline Plus with data regions (in-region processing and storage), DLP/IRM, and Client-side encryption (CSE) used to keep sensitive content out of Gemini.
Only for content you protect with CSE. Google's Privacy Hub says CSE "can restrict Gemini's access to sensitive data, because no Google system or Google employee have the technical means to access CSE content". That statement covers CSE-protected Drive, Docs, Gmail, Calendar and Meet content only, and it also means no Gemini features on that content. Everything else Gemini retrieves uses Google-managed keys, and Google documents no customer-managed key option for Gemini interactions.
Google's Privacy Hub says: "Your content is not human reviewed or otherwise used for Generative AI model training outside your domain without permission." Third-party support subprocessors can see Customer Data only if you grant access in a support case, and a Security Risk Detection activity allows limited human review of accounts suspected of compromise or abuse. Google doesn't state whether Access Transparency or Access Approval logs cover Gemini interactions.
Not without your permission. Google says: "Workspace does not use customer data for training models without customer's prior permission or instruction." That's backed by the Training Restriction in the Workspace Service Specific Terms. Voluntary in-product feedback can include the prompt, source documents or emails and the output; it's kept up to 18 months, and Google says it isn't used to train the models behind Workspace generative AI services. The hub was updated 2026-07-01 for new feedback and data donation flows.
VendorScore doesn't certify compliance; Google publishes a DPA at https://workspace.google.com/terms/dpa_terms.html. The Cloud Data Processing Addendum gives at least 30 days' notice before a new subprocessor processes Customer Data; you can object by terminating within 90 days.
Google's Privacy Hub lists no third-party AI model provider: Gemini runs on Google's own models. The third-party subprocessors for Workspace Core Services are technical-support vendors (for example Accenture, Cognizant, Deloitte, EPAM, GlobalLogic and Tata Consultancy Services), which can access stored Customer Data only if you grant it in a support case. The list is published at https://workspace.google.com/terms/subprocessors/.
Yes, on the right plan. Data regions cover Gemini app and Google Workspace with Gemini prompts and responses at rest and in processing (Europe or US). In-region processing needs Enterprise Plus or Frontline Plus, and Gemini Notebook is excluded from data-region settings.
Every URL cited on this page. Score, key findings and FAQ: VendorScore report (gemini-workspace.md), evidence dated 2026-10-07.