Video Visual China Group | Getty Images
google is letting users test a new artificial intelligence tool that connects information from different apps like Gmail and Google Photos to provide personalized answers with a chatbot.
Google said in a blog post Wednesday that the feature, called Personal Intelligence, will be available for personal accounts.
“Whether it’s connecting an email thread to a video you watched or finding subtle nuances in your photo library, Gemini can now understand context without being told where to look,” Josh Woodward, vice president of Google Labs and Gemini apps, wrote in a post.
This is Google’s latest attempt to build more inference AI capabilities within its Gemini app and attract users as the company takes on the likes of OpenAI in the increasingly competitive generative AI market. Gemini can already pull information from Google apps, but the company announced Wednesday that Gemini 3 can now “reason with your data to uncover proactive insights.”
The new feature conflicts with apple Intelligence is Apple’s personal AI system that integrates apps that help you write, create images, and understand context. Apple announced Monday that it has selected Google to enhance its AI capabilities, including major upgrades to Siri expected later this year.
Personal Intelligence will first be available to Google AI Pro and AI Ultra subscribers in the US, and will be added to Google’s AI Mode search tool in the future, the company said. Google says the feature will be turned off by default.
Woodward, who has launched successful products with Gemini over the past year, warns that beta versions of new features are not free from mistakes and is asking users for feedback.
“Gemini may also struggle with timing and nuance, especially when it comes to relationship changes such as divorce, or different interests,” he writes. When it comes to sensitive topics like health, “Gemini aims to avoid making aggressive assumptions,” but added, “We are willing to discuss this data if asked.”
Woodward also said that the company doesn’t train its AI models “directly on your Gmail inbox or Google Photos library,” but uses “limited information, such as Gemini’s specific prompts and the model’s responses,” to improve its functionality over time.
Featured: Google unveils ‘Gemini 3 Flash’ AI model focused on speed and cost
