Apple Adds Privacy Alerts for AI Features Using Google Cloud Servers
In Focus
- Permission alerts notify users when their data is processed on Google Cloud servers
- The alerts are already displaying on Shape Generation tool in iWork
- Apple routes requests that its servers cannot process to Google Cloud
Apple has started showing permission alerts when certain AI tasks send user requests to Google Cloud. The AI-powered Shape Generation tool in iWork on iOS 26, and similar features in Freeform on iOS 27 already display these alerts.
The alerts notify users when Apple AI tools process requests on Google Cloud servers. Apple places Google’s infrastructure under the same privacy framework that applies to third-party AI services such as ChatGPT.
Why is Apple Introducing Permission Alerts to AI Features?
Apple’s AI permission alerts are part of a shift in the company’s broader AI strategy. In June, the tech giant announced plans to power its next-generation Apple Intelligence with custom models trained in partnership with Google. The models would run on the Google cloud infrastructure, supported by the Gemini architecture.
Apple’s AI partnership with Google has been structured in a way that maintains clear boundaries and includes powering Siri with Gemini. However, the iPhone maker does not use Gemini models that Google has deployed to its customers. The company does not use Google’s client-facing code or Search infrastructure as its knowledge source.
Instead, Apple developed its own foundation models inspired by Gemini’s architecture. An orchestration system directs each request to the most suitable model based on complexity and context required.
How the Permission Alerts Update Improve Apple AI Privacy
Apple’s permission prompts enable the tech giant to strike a balance between processing user data on the device and sending it to third-party servers. On-device processing poses the challenge of limited capability while third-party processing risks user privacy.
Apple sends complex or context-dependent requests to its Private Cloud Compute servers. The company routes requests that its servers cannot process to Google Cloud. To secure the data processed on Google’s systems, Apple relies on Nvidia’s confidential computing technology.
This technology keeps data and AI models encrypted during processing, a move that blocks access even from Google staff. This encryption comes with a performance cost and the query processing process is slightly slower. However, Apple considers this a reasonable tradeoff for maintaining privacy standards across the Google cloud infrastructure.
What the New Permission Alerts Mean for Apple
The new permission prompts reflect Apple’s efforts towards making AI data flows more transparent instead of handling them silently in the background. The move varies from the company’s initial Private Cloud Compute approach where user alerts were unnecessary because data stayed within Apple’s ecosystem.
It’s still early to know whether users will view this transparency as a privacy advantage or a sign of greater reliance on third-party infrastructure. But even as Apple applies encryption protections, the use of external cloud systems could challenge its long-standing privacy-focused image.
