Apple’s AI pitch will live or die by its privacy promise

This review of The Verge focuses on Apple's core statement on AI at WWDC 2026: Apple is not rushing to catch the generative AI boom, but is spending time making AI more "right," with privacy being the most important definition. The article points out that the new Apple Intelligence and Siri AI will span iPhone, iPad, Mac, Apple Watch, and Vision Pro, including a ChatGPT-like Siri AI app, AI camera and photo editing features, and an initial agentic experience that enables Siri AI to interact with other apps and software. Apple's architecture allows processing on-device to remain locally, and only sends to Private Cloud Compute when processing is impossible; The official statement states that the data will not be stored, will only be used to fulfill requests, and will not be accessed by Apple or others. Conversation history in the Siri AI app is kept on the device and in end-to-end encrypted iCloud. The article also reminds us that Private Cloud Compute is not a new concept; it was introduced back when Apple Intelligence first launched in 2024; But two years later, the situation is different. Apple still lags behind most competitors in AI capabilities, so privacy is even more necessary to differentiate itself. At the same time, Apple is catching up by collaborating with Google, Nvidia, and others: the new cloud AI model is based on Google Gemini, and Private Cloud Compute has expanded to Google Cloud, using Nvidia GPUs, Intel CPUs, and Google Titan chips. This weakens the original narrative of "complete control by Apple Silicon and the Apple supply chain," although Apple claims it will manage Google Cloud hardware with a cryptically authenticated, tamper-proof hardware ledger while retaining full software control. The article argues that compared to Google Gemini, OpenAI ChatGPT, and Anthropic Claude, which typically collect or retain more prompts, content, device information, and training data by default, Apple can maintain less default data collection and still maintain credible differentiation. However, whether this selling point can be validated depends on whether users believe that the expanded Private Cloud Compute can still uphold its original security and privacy commitments.