Google's Gemini now retains your conversations for years

A new support page update from Google highlights how it gets data from users of its Gemini chatbot apps for iOS, Android, and the web.

Although talks with Gemini are "disconnected" from Google Accounts, the company emphasizes that human annotators regularly read, classify, and process the exchanges to enhance the service. Google doesn't specify if these annotators are in-house or outsourced, which may be important regarding data security. These discussions, together with "related data" such as the user's location and the languages and devices they used, are stored for a maximum of three years.

The company now gives customers some control over what Gemini-relevant data is kept around and for how long.

Turning off Gemini Apps Activity in Google's My Activity dashboard (which is on by default) will stop any further talks with Gemini from being saved to a Google Account for review, so the three-year window will not apply. On the Gemini Apps Activity panel, prompts and discussions with Gemini can be removed individually.

However, to "maintain the safety and security of Gemini apps and improve Gemini apps," Google states that Gemini discussions will be kept to a Google Account for up to 72 hours even when Gemini Apps Activity is turned off.

It also notes that it is important that users refrain from entering confidential information in their conversations or any other type of data that they wouldn't want reviewers to see or used by the company to enhance its machine learning systems.

In all fairness, the company's policies on the gathering and keeping of GenAI data don't really differ from those of its competitors. For example, OpenAI retains all of its talks with ChatGPT for 30 days, unless the user has paid to an enterprise-level plan with a specific data retention policy, in which case ChatGPT's conversation history function is not saved.

However, Google's policy highlights the difficulties in striking a balance between privacy and the development of GenAI models that use user data to learn and adapt.

Recently, vendors have found themselves in hot water with regulators due to their GenAI data retention practices. The Federal Trade Commission (FTC) asked OpenAI to provide comprehensive information last summer on how it verifies data—including customer data—used to train its algorithms and how it safeguards that data from unauthorized access. 

Organizations are becoming more cautious about privacy issues as GenAI products become more widespread.

According to a recent Cisco poll, 27% of businesses have prohibited GenAI, while 63% of businesses have placed restrictions on the types of data that can be entered into GenAI platforms. According to the same survey, 45% of workers have fed "problematic" data—such as employee records and confidential company information—into GenAI technologies.