Google’s new Personal Intelligence, a new capability for its Gemini artificial intelligence (AI) assistant, can access and analyze data across users’ connected Google accounts, marking the latest development in the intensifying competition among tech giants to deploy AI agents with broader system access.
The feature lets Gemini pull information from Gmail, Google Drive, Maps, and other Google services to provide more personalized responses. However, the launch comes as the AI industry grapples with mounting concerns about data privacy and security as assistants gain deeper access to users’ digital lives.
Google’s move follows similar moves by competitors. Microsoft Corp. has been enhancing its Copilot platform with long-term memory capabilities and integration with third-party services including Google Drive and Gmail. The company also introduced Actions, which allow the assistant to perform tasks such as booking tickets and making reservations autonomously. Meanwhile, Anthropic recently launched Claude Cowork, a file-managing AI agent targeting non-technical users that works proactively within their files.
Expanded capabilities of these AI assistants raise significant privacy and security questions. Connected accounts could potentially expose sensitive information including personal details from email correspondence, financial data from banking notifications, and location histories derived from photos and Maps activity. Security experts warn that if a user’s Google account is compromised, unauthorized access could grant bad actors entry to information aggregated from multiple services.
The introduction of Personal Intelligence reflects the tech industry’s push toward more autonomous and contextually aware AI assistants. However, it also highlights the delicate balance companies must strike between delivering powerful, personalized experiences and protecting user privacy. As AI agents become more deeply integrated into users’ digital ecosystems, how tech companies handle data access and security will likely face increasing scrutiny from regulators and privacy advocates.
Acknowledging concerns, Google has implemented several safeguards for Personal Intelligence. The feature will be disabled by default, requiring users to actively opt in and connect their apps to Gemini. Even after activation, Google says the system won’t be used for every query, but only when Gemini determines that accessing personal data would enhance the response.
The company has also built additional protections for sensitive information. Gemini has been trained to avoid making proactive assumptions about personal health details and other sensitive data. Google is encouraging users to provide feedback about the personalization feature through the thumbs-down button, allowing the company to refine the system’s behavior based on user concerns.
Google has not yet announced a timeline for broader availability of Personal Intelligence beyond its initial rollout.
