AI is slowly infiltrating every part of our daily lives, especially in the IT world. Google is among the leaders in this space and the company has been trying to get customers to use its Gemini model in various products like Workspace and Translate. Now, it is convincing its user base to leverage Gemini while using Gmail too.
Google has been integrating Gemini into Gmail for quite a while now. Its AI model can surface AI Overviews, help you write emails, suggest replies, proofread your content, prioritize important emails, and summarize your inbox, among lots of other things. In its latest blog post, the tech giant is now emphasizing that your Gmail data is completely secure and private even when you use it with Gemini.
Google says that it does not train its "foundational" AI models, such as Gemini, on your personal emails. This is because Gmail is secure by design and any tasks that you delegate to Gemini are performed in an isolated environment.
Secondly, any data used by Gemini is ephemeral by nature in terms of access. The AI model loses access to your inbox" data as soon as it performs the task you asked it to do, so there is no chance of data leakage.
Gmail"s VP of product, Blake Barnes, likens the process to Gemini entering a "private room" which contains all your private data which is leveraged by the AI model to perfom a task, after which it leaves the room and loses access to the repository of information. Check out the YouTube Short on the topic below:
Looking at some historical data where there have been instances of data being leaked by AI models, it"s not entirely surprising that Google wants to emphasize the idea that its different than its competitors. For example, Microsoft was caught accidentally uploading confidential emails to Copilot for summarization purposes. As such, it"s understandable that Google doesn"t want to be categorized in the same space, and is encouraging its customers to give Gemini a try in Gmail, without worrying about privacy or security.