As Apple prepares for its upcoming WWDC event, new insights reveal the company's ongoing efforts to enhance its AI capabilities through a collaboration with Google. The report indicates that a significant focus will be on maintaining on-device processing for AI features, a priority for Apple moving forward.
According to The Information, Apple is utilizing a modified version of Google’s large Gemini model to create a smaller variant that can operate directly on Apple devices. This method, referred to as distillation, aims to optimize AI performance while retaining user privacy.
Additionally, Apple is exploring potential acquisitions of smaller firms, such as Liquid AI, which specializes in local AI processing. Despite these advancements, many AI tasks will likely still rely on cloud support due to the complexity of the full Gemini model, which contains trillions of parameters.
The report also highlights Apple's recent approval of a privacy technology from Nvidia, indicating that some queries will be processed in Google Cloud using Nvidia’s AI chips. This decision, made in the past few weeks, reflects Apple's commitment to balancing cloud capabilities with stringent privacy standards.