Don't enter anything criminal or anything you don't want others to see into Gemini, Google's GenAI family of apps.
That's the PSA (of sorts) released by Google today, outlining in a new support document how the company collects data from users of its Gemini chatbot app for the web, Android, and iOS.
Google says human annotators routinely read, label, and process your conversations with Gemini to improve its services (even though the conversations are “disconnected” from your Google account). ). (It's not clear whether these annotators are in-house or outsourced, but that could be important when it comes to data security; Google hasn't said so.) It will be stored for up to 3 years along with other related data. Devices used by users and their locations.
Currently, Google offers users Several Control what and how Gemini-related data is retained.
Google's[マイ アクティビティ]If you turn off Gemini app activity in your dashboard (enabled by default), future conversations with Gemini will no longer be saved to your Google Account for review (i.e. the 3-year period will not apply). yeah). Meanwhile, individual prompts and conversations with Gemini can be deleted from the Gemini app's activity screen.
But Google says that even when Gemini app activity is turned off, Gemini conversations will be stored in your Google account for up to 72 hours to “maintain the safety and security of the Gemini app and improve the Gemini app.” ing.
“Don't enter confidential information in your conversations, data you don't want reviewers to see, or data you don't want Google to use to improve our products, services, or machine learning technology,” Google wrote. ing.
To be fair, Google's GenAI data collection and retention policies aren't all that different from those of its competitors. For example, OpenAI retains 30% of all chats with ChatGPT, regardless of whether ChatGPT's conversation history feature is turned off, unless the user subscribes to an enterprise-level plan with a custom data retention policy. Save for days.
But Google's policy illustrates the challenges inherent in balancing privacy with the development of GenAI models that feed user data for self-improvement.
GenAI's liberal data retention policy recently led the vendor to compromise with regulators.
Last summer, the FTC asked OpenAI to explain how the company vets data used to train its models, including consumer data, and how that data is treated when accessed by third parties. They requested detailed information on whether they were protected. Overseas, Italy's data privacy regulator, the Italian Data Protection Agency, said OpenAI had no “legal basis” for the mass collection and storage of personal data to train GenAI models.
As GenAI tools proliferate, organizations are becoming increasingly wary of privacy risks.
A recent Cisco study found that 63% of companies place limits on the data they can input into their GenAI tools. Meanwhile, 27% have completely banned his GenAI. The same study revealed that 45% of employees enter “problematic” data into his GenAI tool, such as employee information or private files about their employer.
OpenAI, Microsoft, Amazon, Google, and others offer GenAI products explicitly for enterprises such as: please do not Retain data for any length of time, even for model training or other purposes. But as is often the case, consumers are left in the lurch.