Apple Preps "Core AI" to Replace Core ML

The branding shift from "machine learning" to "artificial intelligence" is a deliberate move to modernize the framework and better resonate with developers and consumers. This change is expected to be a central pillar of the upcoming iOS 27, signaling a deeper integration of AI across all of Apple's platforms. Core ML, first introduced in 2017, established Apple's long-standing strategy of prioritizing on-device processing for privacy and performance. While Core ML received updates for on-device training and support for generative models at WWDC 2025, Core AI represents a more foundational overhaul rather than an incremental update. This architectural change is driven by a need to simplify how developers integrate external AI models. The move aligns with CEO Tim Cook’s multi-model strategy for Apple Intelligence, which includes leveraging Google's Gemini for some features and an existing partnership with OpenAI. The strategic pivot in software follows a significant leadership transition in late 2025. John Giannandrea, who was hired from Google in 2018 to lead AI, announced his retirement and was succeeded by Amar Subramanya, a former AI executive from Google and Microsoft. Under the new structure, Subramanya reports directly to Craig Federighi, Apple's Senior Vice President of Software Engineering. This reporting line suggests a tighter integration of AI development within the core OS teams, a departure from the previous structure. Core AI will serve as the developer-facing component for Apple's broader strategy, which utilizes a ~3 billion parameter on-device model for speed and privacy, alongside larger server-based models running on Apple silicon in its Private Cloud Compute environment for more complex tasks. To ensure a smooth transition, the Core ML framework may coexist with Core AI for some time. The ultimate goal is to provide developers with a more robust, unified set of tools that reduce the need to build AI features from scratch.