Apple’s AI Revolution: Enhancing iOS 18 with In-House Language Models and Offline Functionality

Apple has announced plans to focus on enhancing the Artificial Intelligence (AI) features of iOS 18 by integrating a large language model (LLM) developed in-house into iPhone smartphones. This will enable AI processing to happen on the device itself, without needing to access the cloud, prioritizing privacy. The company is also working on new AI-based features in preparation for the upcoming iOS 18 operating system update.

One of these new features is integration with Google’s generative AI engine, Gemini, through software updates later this year. Apple’s focus is on introducing improvements that enhance functions directly on its devices, rather than relying on cloud-based services. This includes AI capabilities to create images or write essays based on simple instructions from the device.

Apple researchers have developed a language model called RealLM (Reference Resolution as Modeling Language) that enhances the naturalness of conversation by enabling AI to understand ambiguous references in human language. While integrated LLM devices may have fewer training parameters than cloud-based models, Apple continues to work on new tools and technologies to implement AI functions directly from iPhones.

In addition, Apple is working on a ‘chatbot’ named Ajax to challenge other companies in the AI technology space. The company aims to boost iOS 18’s AI features with a self-developed LLM model that can function seamlessly on the device itself, enabling offline AI functions to be performed securely and privately without relying on external connections.

At the upcoming Worldwide Developers Conference (WWDC) 2024, which will take place from June 10 to 14, Apple is expected to announce more news about its AI and language models development efforts. The company is actively investing in this area and aims to improve the capabilities of its devices through cutting-edge technology.