In recent updates, Apple indirectly confirmed that certain AI functions in iOS 18 and other Apple operating systems will operate without an internet connection. This confirmation came alongside the introduction of the OpenELM family of large language models (LLMs), which can run locally on devices.
OpenELM Models: Enabling Offline AI Functions
Apple’s OpenELM models, now available as open source through the Hugging Face community, offer a crucial advancement in AI technology. These models facilitate the generation and processing of text directly on Apple devices, eliminating the need for remote server access.
At present, the OpenELM library encompasses eight models, including both pre-trained and configured versions with parameters ranging from 270 million to 3 billion. Third-party developers have immediate access to these models for commercial utilization and customization, marking a significant stride towards democratizing AI development.
Implications and Future Prospects
Apple’s decision to make OpenELM accessible to third-party developers suggests a broader strategy to integrate AI technologies not only into its own products but also across the developer ecosystem. This approach mirrors the successful adoption of augmented reality technologies, hinting at a potential shift in the AI landscape within the Apple ecosystem, notes NIX Solutions.
We anticipate further insights into Apple’s AI roadmap and the implications of OpenELM at the WWDC 2024 conference, scheduled for June 10. As developments unfold, we’ll keep you updated on the advancements and potential applications of offline AI functions in Apple’s operating systems.