CUPERTINO, Calif. — Apple used its spring event Tuesday to lay out the most detailed version yet of its artificial-intelligence strategy, pairing a more capable Siri with generative tools designed to run directly on iPhones, iPads and Macs whenever possible.
The announcement matters because Apple has been under pressure to show how it will compete in a market reshaped by OpenAI, Google, Microsoft and Meta. Instead of framing AI as a standalone chatbot, Apple presented it as a system-level layer that can summarize information, understand context and complete tasks across apps.
Chief Executive Tim Cook said the company's goal is to make AI feel personal without making users give up privacy. Apple repeatedly emphasized that many requests would be processed on device, while more complex tasks could be routed through protected cloud infrastructure or optional third-party models.
The most visible update is Siri. Apple said the assistant can now maintain context across follow-up questions, understand what is on screen, search personal data with permission and take multi-step actions inside supported apps. In practical terms, that could mean finding a document, summarizing it and attaching it to an email without the user jumping between apps.
Apple also introduced writing tools, notification summaries, image-generation features and developer APIs that allow apps to expose actions to the assistant. The company said developers will be able to define what Siri can do inside their apps while preserving user consent and data controls.
Privacy was the central selling point. Apple said simple requests can stay on the device, while cloud requests will use a system that limits data retention and makes server behavior auditable. That approach is meant to reassure users who want AI convenience but are wary of sending personal messages, photos or calendar data to remote models.
The strategy carries trade-offs. On-device models may be faster and more private, but they can be less capable than the largest cloud systems. Apple appears to be betting that users will prefer reliable everyday assistance over the most experimental model behavior.
For developers, the event opens a new competitive field. Apps that integrate cleanly with Apple's AI layer may become easier to use, while apps that do not could feel disconnected from the operating system. That makes the new APIs as important as the consumer-facing demos.
The rollout will also test Apple's ability to explain AI features clearly. Consumers have heard broad promises about artificial intelligence for years; what matters now is whether the tools save time in common tasks such as messaging, search, scheduling, photo management and accessibility.
Apple did not end the AI race with one keynote. But it clarified its lane: AI built into the devices people already use, with privacy and practical utility as the pitch. The next question is whether the features work consistently enough to become daily habits.
Priya Narayanan
Technology editor and former Silicon Valley software engineer.
