To stop Google Gemini's 'help me write' pop-ups in Google Docs in 2026, users must navigate specific document menus or disable all 'smart features' across their entire Google Workspace, including Gmail. An all-or-nothing privacy trade-off is presented. While Google offers document-level controls, a complete AI opt-out requires sacrificing other convenient functionalities and granting AI unfettered access to private information. This lack of granular control will likely fuel user frustration and demand for more flexible, privacy-respecting options.
Disabling AI Suggestions in Google Docs
Google Docs features a 'help me write' function, often hovering over the cursor, providing AI-generated suggestions. To turn off these 'write with Gemini' pop-ups, users click 'Gemini' on the top menu bar, select 'bottom bar preferences,' then choose to turn off the bottom bar, according to TechCrunch. Alternatively, Gemini suggestions can be disabled at the document level by clicking the Gemini icon in the toolbar and selecting 'Bottom panel settings,' as detailed by Mezha. These methods offer direct, document-specific control, but Google fails to provide a consistent, system-wide opt-out.
The All-or-Nothing Workspace Control
Disabling Gemini pop-ups across the entire Google Workspace requires navigating to Gmail settings, clicking 'See all settings,' and then 'Manage Workspace smart feature settings,' TechCrunch reports. This forces an all-or-nothing choice: users either accept the full suite of AI functionality or lose all AI features, with no granular control over individual Gemini functions, BGR states. This approach demands users grant AI unfettered access to private Gmail data, a cost many find unacceptable for convenience, BGR notes. Even minimal additional user control over Gemini's Workspace usage would improve the current binary option, BGR suggests.
The Illusion of Granular AI Control
While document-level controls exist, managing Docs AI from Gmail settings reveals a centralized AI strategy. This prioritizes Google's data access over intuitive user control and feature-specific privacy preferences. Such a convoluted path suggests Google intends to deeply embed AI, making selective opt-out difficult and potentially user-hostile, according to TechCrunch. This effectively strong-arms users into broad data sharing, bundling AI features to implicitly condition acceptance of undifferentiated AI data access.
Outlook for Google Workspace AI
If Google fails to offer more granular AI controls, user dissatisfaction and calls for greater data privacy are likely to intensify.
