Apple Explores On-Device AI for iPhones with Startup PrismML

Shrinking AI for Mobile Devices

Apple is reportedly in early discussions with PrismML, a startup specializing in compacting large artificial intelligence models. This technology aims to enable advanced AI functions to run directly on smartphones. The move could significantly reduce Apple's reliance on cloud-based processing for its virtual assistant, Siri.

Why is On-Device AI Important for Apple?

PrismML's chief executive, Babak Hassibi, confirmed these preliminary talks to CNBC. He stated that Apple, among other companies, is evaluating their innovative technology. These discussions are still in their very initial stages.

PrismML's core innovation lies in its ability to condense massive AI models. These models typically require powerful data centers to operate effectively. By shrinking them, PrismML allows complex AI to function locally on devices like i Phones. This on-device processing offers several advantages, including enhanced privacy and faster response times. It also reduces the need for constant internet connectivity for AI tasks.

# What is the main goal of PrismML's technology?

Integrating PrismML's technology could transform how Siri and other Apple AI features operate. Currently, many Siri requests are sent to Apple's cloud servers for processing. Bringing these operations onto the device would improve user privacy, as data would not leave the phone. It would also likely make Siri faster and more reliable, especially in areas with poor internet service. This strategic shift could give Apple a significant edge in the competitive AI landscape.

# How might this technology benefit i Phone users?

PrismML aims to reduce the size of large AI models. This allows them to run efficiently on mobile devices such as smartphones.

Users could experience faster AI responses and improved privacy. Their data would be processed directly on their device, not in the cloud.

# Are these discussions between Apple and PrismML finalized?

No, the discussions are still in their very early stages. No definitive agreements have been made yet.