Apple's new AI model unveils local AI integration for iPhones

Apple is stepping into the realm of generative AI with the launch of its new OpenELM series, signaling a significant shift towards running AI directly on Apple devices.

These lightweight language models, released by Apple researchers on the Hugging Face model library, promise efficient performance on text-related tasks, such as email writing, all while running locally on Apple devices.

Dubbed "Open-source Efficient Language Models," or OpenELM, these models come in four sizes, ranging from 270 million to 3 billion parameters. Parameters represent the variables a model understands in decision-making, and Apple's models are notably smaller than their counterparts from Microsoft and Google. This downsizing ensures cost-effectiveness and optimization for devices like iPhones and laptops.

Apple CEO Tim Cook previously hinted at the integration of generative AI features into Apple devices, emphasizing the company's dedication to this space. However, specific details regarding the application of AI on Apple products remain undisclosed.

While Apple has previously released AI models like MLX for machine learning optimization and MGIE for image editing, the OpenELM series marks a significant step towards local AI integration. Despite these successes, Apple has reportedly reached out to Google and OpenAI, indicating the possibility of a collaboration to further expand artificial intelligence capabilities in Apple products.