Apple has announced that its new "Siri AI" will leverage Google's Gemini models, running on Google's Nvidia hardware. To address privacy concerns, Apple is implementing a new iteration of its Private Cloud Compute system that utilizes third-party hardware, including Nvidia's Confidential Computing and Intel's Trust Domain Extensions. While simpler queries will still be processed on-device or on Apple's own servers, more complex reasoning and agentic tasks will be handled by these secured Google Cloud servers, with Apple maintaining a verifiable ledger of the hardware used and ensuring only Apple-signed software runs on it. AI
IMPACT Apple's approach to AI privacy in cloud processing may set a precedent for other companies integrating third-party AI models.
RANK_REASON Company announcement regarding AI product integration and privacy measures.
Read on Mastodon — mastodon.social →
AI-generated summary · Google Gemini · from 5 sources. How we write summaries →