Apple is working on building artificial intelligence technology directly into its hardware, a new report by Bloomberg reveals. According to a “person familiar with the matter,” Apple is developing a new processor, dubbed the “Apple Neural Engine,” that will be dedicated to AI-related tasks. The goal is to improve how AI tasks such as facial recognition and speech recognition are handled by incorporating it into a specialized hardware chip that would optimize performance as well as providing for more advanced capabilities. Complex AI tasks are currently handled by software processes that use the existing CPU and GPU hardware, requiring the workload to be shared with other tasks being performed on the device. A dedicated module would ensure that AI-related tasks are handled more efficiently both in terms of processing speed and battery consumption.
Apple would not be the first company to develop a dedicated AI processor, however. Qualcomm’s latest Snapdragon chip used in some smartphones includes a dedicated processing module for AI tasks, and Google developed a Tensor Processing Unit (TPU) last year for use in its data centers. However, Apple appears to be taking its ambitions in a slightly different direction, with plans to integrate the chip into many of its hardware devices, including the iPhone and iPad, in line with the company’s approach of handling machine learning tasks on board the device rather than offloading tasks to the cloud as Google does. The chip could potentially power features such as facial recognition in the Photos application as well as parts of speech recognition and predictive typing. Sources suggest that Apple also plans to offer developer access to the chip so that third-party apps can take advantage of offloading AI-related tasks. It’s unclear at this time, however, whether the new component is ready to be included in this year’s devices, or is part of a longer-term plan that Apple is working on.