Craig Federighi discusses Apple’s collaboration with OpenAI and the trajectory of Apple Intelligence in a recent interview.
During the unveiling of iOS 18 and macOS 15 at WWDC 2024, Apple introduced Apple Intelligence, a suite of AI-driven features poised to enhance user experiences. While these innovations are slated for release later this year, Craig Federighi, Apple’s head of software engineering, delved into the implications and aspirations surrounding Apple Intelligence in a conversation with Fast Company.
Federighi shed light on Apple’s alliance with OpenAI, a pivotal component in the development of Apple Intelligence. Among the highlights of this collaboration is the integration of ChatGPT into Siri, Apple’s voice assistant. This integration empowers Siri with advanced language processing capabilities, enabling users to perform tasks like text rephrasing, message summarization, image generation, and even emoji creation. Federighi emphasized the importance of this partnership in broadening the accessibility of cutting-edge AI capabilities within Apple’s ecosystem.
Apple Intelligence harnesses a combination of on-device processing and cloud computing to deliver its features effectively. Federighi acknowledged the existence of other robust language models beyond Apple’s proprietary creations, acknowledging the merits of incorporating such models into Apple’s offerings. Notably, Federighi lauded GPT-4o as one of the premier language models available, underscoring Apple’s motivation to integrate it into Siri.
In a commitment to transparency, Federighi assured users that Siri would seek permission before resorting to external language models like ChatGPT to fulfill queries. Furthermore, Apple aims to expand support for third-party language models, affording users greater flexibility in choosing their preferred linguistic tools.
Navigating the complexities of regulatory landscapes, particularly in markets like China, poses unique challenges for Apple’s AI endeavors. Federighi hinted at ongoing efforts to navigate these hurdles, expressing a desire to introduce Apple Intelligence to Chinese users in due course.
Underpinning Apple’s AI infrastructure is Private Cloud Compute (PCC), a network of AI servers designed to process user requests while safeguarding privacy. Federighi elucidated on the secure handling of data requests, highlighting Apple’s commitment to preserving user privacy through cryptographic measures.
Looking ahead, Federighi envisioned a future where advancements in chip technology could facilitate the execution of larger language models directly on devices. However, he underscored the enduring importance of online models in delivering up-to-date information and acknowledged the necessity of accessing external services to augment device capabilities.
As Apple continues to chart the course for AI innovation, Federighi’s insights offer a glimpse into the company’s ethos of integrating cutting-edge technology with a steadfast commitment to user privacy and experience.
