Google launches Personal Intelligence feature in Gemini app, challenging Apple Intelligence

Google has rolled out an innovative artificial intelligence feature called Personal Intelligence, allowing users to receive tailored responses by integrating data from various applications like Gmail and Google Photos. Announced in a blog post on Wednesday, this feature is currently available for personal accounts. Josh Woodward, the Vice President of Google Labs and the Gemini app, emphasized the transformation this technology brings. "Gemini now understands context without needing explicit guidance, whether it's linking an email thread to a video you viewed or extracting details from your photo library," he stated. This launch marks Google's ongoing effort to enhance the reasoning capabilities of the Gemini app as it competes in the burgeoning generative AI landscape, particularly against rivals like OpenAI. Although Gemini previously had the ability to pull information from Google apps, the newly unveiled Gemini 3 enhances this with its ability to reason across user data and provide proactive insights. The Personal Intelligence feature is set to compete directly with Apple Intelligence, Apple's own integrated personal AI system that assists with tasks like writing and image generation. Recently, Apple announced that it has chosen Google to enhance its AI functionalities, including a significant upgrade to Siri anticipated later this year. Initially, Personal Intelligence will be available to Google AI Pro and AI Ultra subscribers in the United States, with plans to incorporate it into Google's search tool's "AI Mode" in the future. By default, the feature will be disabled, ensuring users can opt-in based on their comfort level. Woodward acknowledged that, as with any beta version, some errors may still occur and invited users to provide feedback to refine the tool. He noted that Gemini may have difficulties with timing or nuances, especially concerning sensitive topics like personal relationships or health matters. The AI is designed to avoid making assumptions but will engage in discussions about such topics if prompted. Additionally, Woodward clarified that Google does not train its AI models using data directly from users' Gmail or Google Photos, but rather employs limited information, such as specific prompts and the AI’s responses, to enhance the model's performance over time.