Skip to content
Apixo
Blog
news· 2 min read· via ZDNet AI

Google Updates Gemini to Separate Chat History From AI Training

Google is updating its Gemini settings to let users opt out of AI model training without losing their chat history, aligning with industry standards.

Google Updates Gemini to Separate Chat History From AI Training

Google is rolling out changes that will make it much simpler for users to opt out of having their interactions used for Gemini model training. Previously, users who had privacy concerns faced an all-or-nothing choice, as preventing Google from using chats for training meant deleting their entire conversation history.

Set to arrive in the coming weeks, this update removes that limitation. Google announced via email to users that the overhaul will decouple conversation retention from machine learning training efforts.

A Closer Look at the New Settings

Until now, Gemini relied on a single setting named Keep Activity. This control managed both how long conversations were saved and whether those dialogues contributed to AI improvement, combining two distinct preferences into one.

Google is replacing Keep Activity with two separate controls. The first setting handles chat history retention, allowing users to keep or delete past conversations and select how long records are stored. The second setting focuses purely on model training, giving individuals the choice to permit or block their texts and attachments from being utilized for AI development.

This adjustment aligns Google with other major artificial intelligence providers, which already allowed users to keep historical records while opting out of data harvesting for training.

Industry Standards for Privacy

Most conversational AI platforms feature manual toggles for training data, but generally do not force history deletion as a penalty for opting out. ChatGPT uses an option called “Improve the model for everyone,” Microsoft Copilot provides specific switches for conversation and voice training activities, and Claude disables AI training by default while offering a “Help improve our AI models” checkbox.

Google was previously the sole major AI provider enforcing the restriction that tied training consent directly to history retention, a policy that frequently drew criticism regarding user privacy.

What it means for developers

Developers building applications or managing data pipelines should note how major platforms handle user privacy and data governance. Managing user consent cleanly is critical when integrating large language models. For those building and testing applications, developers can try top AI models cheaply through one API at https://apixoai.online to compare behaviors and integrate multiple systems efficiently.

While Google’s update primarily affects consumer-facing accounts, it marks a notable shift toward standard privacy practices across the broader artificial intelligence landscape.


Source: Google is making it easier to opt out of Gemini AI model training – here’s how — ZDNet AI. Written by the Apixo team from that report.

#ai-news#google#gemini#ai#privacy#developers
Try it with your own tools

One key for Claude, GPT, GLM, DeepSeek and more. Pay per token with crypto.

Get your API key

Keep reading