ALBAWABA – As Apple gears up to announce its anticipated iOS 18 at the 2024 Worldwide Developers Conference at upcoming June, expected to be the iPhone’s biggest update so far with a medley of advanced AI features, it seems that the tech giant is taking a different path with its upcoming Large Language Model.
A Large Language Model, also known as an LLM, is a kind of deep learning algorithm that is capable of performing a wide range of natural language processing tasks and is trained using enormous datasets to recognize, translate, forecast, or create text or other material, an example of those would be ChatGPT, Google Gemini and Meta AI.
Apple is reportedly working on its own LLM, which is expected to power AI features in the upcoming iOS 18, however, according to Mark Gurman from Bloomberg, Apple is planning to run the AI model locally on-device, rather than through a cloud service like other providers.
A new report says #Apple’s AI features coming to iOS18 will run on-device and not on a server, protecting user privacy. pic.twitter.com/cZoWTLg23H
— AppleInsider (@appleinsider) April 16, 2024
This means that instead of sending the prompt to a far-away server that processes the user’s query and then sends back an answer across the internet, the processing will be handled on the user’s own iPhone directly, which wccfetch expects to allow the LLM model to respond in real-time.
Apple's direct cloud-based competitors may have more powerful AI capabilities than Apple in certain cases due to having limited processing power on the iPhone compared to big servers, as MacRumors notes, but Gurman said that Apple might "fill in the gaps" by borrowing technology from Google and other AI service providers like OpenAI to support its model.
In February, Apple CEO Tim Cook stated Apple is spending a "tremendous amount of time and effort" on artificial intelligence and would reveal its results "later this year," with MacRumors reporting that Apple is committing over $1 billion annually in AI research and the technology required to operate complex language models.
