Categories
Technology

Google updates Gemini chat history

Users can now save Gemini chats without allowing Google to train AI models

Google is changing the way its Gemini artificial intelligence (AI) chatbot handles user conversations, introducing separate privacy controls for chat history and AI model training. The move will allow users to save and revisit their conversations without automatically allowing Google to use the content to improve its generative AI models.

Announced on November 16, the update is aimed at giving Gemini users greater control over their personal information. Until now, Google’s ‘Keep Activity’ setting governed both the storage of conversations and the use of chat data for AI training. Users who wanted to retain their chat history had limited flexibility over how their information was used.

With the new arrangement, Google is separating these two functions. Users will be able to keep their Gemini chat history switched on while independently opting out of allowing their conversations to be used for AI model training. The changes will begin rolling out globally through the Gemini app over the coming weeks.

The update is significant for people who regularly use AI chatbots for personal and professional tasks. Many users rely on Gemini to draft emails, summarise documents, conduct research and organise their daily activities. While having access to previous conversations is useful, concerns over how personal information is collected and used have made privacy an important consideration.

Gemini gets separate history and training settings

Under the revised system, the History setting will determine whether conversations are saved and how long users can access them. A separate control will govern whether chat content can be used to train Google’s generative AI models.

This means users will no longer have to disable their entire chat history to prevent their conversations from being used for AI training.

The training setting covers a wide range of information shared through Gemini. This includes text prompts, audio transcripts, photographs, documents, videos, screen content, responses generated by the chatbot, tasks and interactions with connected tools and applications.

When users switch off the AI training option, their future conversations will not be used to train Google’s generative AI models unless they voluntarily submit feedback. However, this does not mean Google stops processing all information associated with those interactions.

The company may continue using relevant conversations to maintain the security of its services, investigate potential problems and improve the overall user experience.

Google says conversations selected for review are separated from users’ Google accounts before being shared with service providers. This is intended to provide an additional layer of privacy during the review process.

Temporary chats will be retained for two months

One of the more notable aspects of the update concerns how Google handles conversations when users switch off their chat history.

Previously, Gemini conversations created with Keep Activity disabled were generally retained for around 72 hours. Under the revised policy, Google will retain future chats for approximately two months, even when the new History setting is switched off.

The company says the extended retention period is necessary to support safety and security measures, respond to user feedback and protect Google, its users and the wider public.

This means disabling chat history will no longer result in conversations being removed from Google’s systems after just a few days. Instead, the information will remain stored for a longer period for operational and security purposes.

Google has clarified that conversations created when History is switched off will not be used to train its generative AI models, unless users submit feedback. The company may nevertheless use the information to respond to requests and investigate issues.

The distinction is important because turning off chat history should not be confused with immediately deleting conversations or preventing all forms of data processing.

What happens when users submit feedback?

Google’s updated privacy policy also explains how information is handled when users voluntarily share feedback about Gemini’s responses.

When feedback is submitted, the company may collect the relevant conversation details, including the last 24 hours of chats, attachments and information from connected applications. These interactions may be reviewed by specially trained human reviewers to identify problems, investigate potentially unsafe responses and improve Google’s services.

According to Google’s privacy documentation, conversations and related information reviewed by service providers may be retained for up to three years, even if users subsequently delete their activity or switch off the relevant settings.

This means users should be aware that changing their privacy preferences does not necessarily erase information that has already been selected for review.

Existing users can manage their preferences

The new controls will be introduced gradually, meaning they may not appear for all Gemini users immediately.

Google says users who had previously switched off Keep Activity will have both new settings disabled by default. Those who had the earlier setting enabled will be able to manage chat history and AI training separately under the revised arrangement.

Users who keep their chat history enabled can also decide how long their conversations remain available. Google’s existing automatic deletion options include three months, 18 months and 36 months, while users can also choose to retain their history until they delete it manually.

The company says users can review or delete their saved conversations whenever they want.

The latest changes come amid growing scrutiny of how technology companies collect personal information and use it to develop artificial intelligence products. As AI chatbots become part of everyday digital activities, the amount of information users share with these platforms continues to increase.

For Google, separating Gemini chat history from AI training is an attempt to address one of the central concerns surrounding consumer AI tools: users want the convenience of personalised and accessible conversations without necessarily allowing their information to contribute to the development of AI models.

While the update offers greater flexibility, the extension of temporary chat retention from 72 hours to two months also raises questions about how long user information remains within Google’s systems.

Ultimately, the changes give Gemini users more choices, but understanding the difference between chat storage, data retention and AI training will remain essential for anyone concerned about online privacy.