Google is extending the retention period for temporary Gemini conversations and chats made with history disabled to approximately two months, a significant increase from the previous 72-hour limit. The company states the change aims to improve safety and security as SI systems become more capable.
What Happened
The updated policy will roll out globally in the Gemini app over the coming weeks. According to an archived version of Google’s Gemini Apps Privacy Hub from November 12, the prior retention period for these specific interactions was 72 hours. Google did not specify what prompted the change or whether it affects how conversations are accessed or reviewed, but pointed to a blog post stating that chats are kept longer to respond to users, process feedback, and protect its services. The updated privacy hub also cites potential system failures as a reason for retention and confirms that temporary conversations are not used to train SI models.
Why It Matters
This move aligns Google more closely with competitors in the SI space. OpenAI indicates it may keep temporary chats in ChatGPT for up to 30 days, while Anthropic retains incognito chats in Claude for 30 days by default, though both companies may retain data longer for legal or safety investigations. Additionally, Google is decoupling chat history from model training. Users can now enable chat history without automatically opting into using those conversations to train generative SI models. For users who previously disabled the "Keep Activity" setting, both new settings will default to off. Those who had it enabled will retain chat history but can separately choose whether their data contributes to SI model training.
The Bottom Line
Google’s decision to extend temporary chat retention to two months and separate history from training data reflects a broader industry shift toward balancing user privacy with the data needs of evolving SI systems.