Gemini’s Personal Intelligence Goes Global—but Access Still Has Gaps
Gemini’s Personal Intelligence has moved beyond the limited U.S. beta introduced in early 2026. As of August 14, the feature is available to eligible personal-account users in more markets, although regional exclusions, gradual activation and service-specific requirements mean it is not universally accessible.
The central capability remains the same: with permission, Gemini can combine information from connected Google services to produce answers shaped by a user’s emails, files, photos, searches and viewing history. That does not mean Gemini automatically “knows everything” about a person; what it can use depends on the account, enabled settings, connected apps and information available in those services.
What changed after the original U.S. beta
Personal Intelligence began as an opt-in beta for eligible Google AI Pro and Ultra subscribers in the United States. Its initial purpose was to move beyond retrieving one email or image at a time: Gemini could reason across several connected services and use the relationship between those records to answer a request.
The rollout subsequently widened. An April 2026 Gemini update announced international expansion for Google AI plan subscribers while excluding the European Economic Area, Switzerland, the United Kingdom, South Korea, Australia and Nigeria at that stage.
This expansion is the important change since the launch. Personal Intelligence is no longer accurately described as a U.S.-only experiment, but “global” should not be read as “available to every Gemini user everywhere.” Eligibility continues to vary by country, account type, device, product surface and connected service.
What Gemini can connect and understand
Gemini personal ai assistant can work with Google Workspace content such as Gmail, Calendar and Drive, as well as Google Photos, Search services, YouTube history and, for some eligible users, Contacts. These connections allow Gemini to retrieve a specific record and combine it with contextual information from another service.
A request about a trip, for example, could draw on an itinerary in Gmail, an event in Calendar and previous destinations represented in Photos. A product-support question could potentially use a purchase receipt to identify the exact device before generating troubleshooting instructions. These are examples of cross-service reasoning, not evidence that Gemini continuously builds a complete or perfectly accurate biography.
The distinction matters because connected information can be incomplete or misleading. A collection of photographs may show where someone has been without explaining why; an old email may no longer describe a current relationship or preference. Personal Intelligence can reduce the amount of context a user must type, but it does not remove the need to check its interpretation.
Eligibility depends on more than location
The current Gemini eligibility and setup guide requires users to be at least 18, sign in with a personal Google Account and keep Gemini Apps Activity enabled. Work, school and supervised accounts are not eligible, and Personal Intelligence remains unavailable in the European Economic Area, Nigeria, Switzerland and the United Kingdom.
Individual data sources add further conditions. Search-based personalization requires Web & App Activity; YouTube-based personalization requires YouTube History; and connecting Workspace content depends on the relevant smart-features setting. Google Photos is supported only in certain countries and requires additional Photos configuration, including Face Groups and identification of the user’s own face group.
Availability also differs across Gemini surfaces. The feature is supported in the Gemini mobile and web apps, Gemini in Chrome where that product is available, and Gemini on compatible smartwatches, but not in every Gemini feature. A person may therefore meet the general account requirements and still lack a particular connection or see the option arrive later through a gradual rollout.
Opt-in control is real, but the data-use details matter
Connecting apps is a choice rather than an automatic consequence of using Gemini. During setup, users can connect all supported services, select individual ones or decline the connections; those choices can later be changed through Personal Intelligence’s Connected Apps settings.
Controls also operate at the conversation level. Personal Intelligence can be disabled for a particular chat, and a personalized answer can be regenerated without personalization. A temporary chat is the more appropriate option when the user also does not want that conversation saved to Gemini Apps Activity.
The privacy claim needs precise wording. Google’s current Connected Apps data-use explanation says its generative models are not trained directly on an entire Gmail inbox, Workspace repository, Contacts collection or the imagery and audio in a Google Photos library; however, when Keep Activity is enabled, relevant summaries, excerpts, generated material and inferences used in a Gemini interaction may contribute to product improvement and model training.
That is narrower than training directly on a complete private library, but it is not equivalent to saying that nothing derived from connected content can enter the improvement process. The same documentation says connected data can include emails, files, events, photos, videos, saved Search and YouTube activity, contact information, location information and insights inferred from those records. Users should evaluate the feature with that broader definition in mind.
How to decide whether the trade-off is worthwhile
Personal Intelligence offers the most value when a request depends on facts scattered across several services: identifying a purchased device, reconstructing travel details, locating a remembered photograph or adapting recommendations to established preferences. For a generic factual question, connecting private services may add little.
A cautious setup is to connect only the service needed for a recurring task and examine how Gemini identifies the information behind its answer. Disconnecting a service prevents future retrieval from that connection, but information already present in saved chats may still affect later personalization; managing activity and connections are therefore separate decisions.
The practical change is reach, not omniscience. Personal Intelligence has progressed from a restricted U.S. beta to a broader international product, while its usefulness still depends on explicit connections and its availability remains uneven. It can assemble unusually personal context, but users retain the responsibility to choose the inputs, verify the output and understand how interaction data may be handled.
Also read:
Subscribe to our newsletter
Get the latest Web3, AI, and crypto news delivered straight to your inbox.