WWDC 2026 TO INTRODUCE CORE AI AS REPLACEMENT FOR CORE ML

Apple is preparing a significant branding and architectural shift in its developer tools, with plans to unveil a new “Core AI” framework at WWDC 2026 that will succeed the longstanding Core ML machine learning framework.

- 

The change, reported by Bloomberg’s Mark Gurman in his Power On newsletter, signals more than a name update — it reflects Apple’s broader repositioning around artificial intelligence across iOS 27 and future hardware. Core ML has powered on-device machine learning in Apple apps since 2017, enabling developers to integrate image recognition, natural language processing, and predictive intelligence directly into their applications. Over time, the framework expanded beyond traditional classification tasks into generative AI and advanced language modeling. Now, Apple appears ready to formalize that evolution under a new identity.

FROM CORE ML TO CORE AI

Core ML originally launched as a privacy-first, on-device framework that allowed developers to deploy trained machine learning models efficiently across iPhone, iPad, Mac, Apple Watch, and Apple TV. It became foundational for features like Face ID, object detection in Photos, handwriting recognition, and Siri’s contextual suggestions.

In recent years, however, the definition of “machine learning” inside Apple's ecosystem has broadened. Generative AI tools, foundation models, and transformer-based architectures have entered the stack. Apple has increasingly referred to its internal “Apple Foundation Models” in developer documentation and sessions.

The anticipated transition from Core ML to Core AI formalizes that pivot. The shift from “ML” to “AI,” as Gurman notes, is symbolic. It aligns Apple's developer language with the industry's broader move toward foundation models, generative agents, and conversational systems rather than narrowly scoped predictive algorithms.

APPLE FOUNDATION MODELS AND GEMINI TRAINING