appleThe highly anticipated iOS 18 and macOS 15 systems were released at WWDC 2024, and one of the important new features is "Apple Intelligence"- a set of artificial intelligence-based tools. The feature will be officially launched later this year. Apple's head of software engineering Craig Federighi revealed some future development plans for Apple Intelligence in an interview with Fast Company.
"Apple Intelligence" will give users a range of practical AI-assisted features, such as helping users rewrite text, summarize information, generate images and even emojis. In addition, thanks to the update of artificial intelligence, Siri will also become smarter, able to understand the context of screen content and allow users to control more aspects of the device.
It is worth mentioning that Apple's "Apple Intelligence" uses a combination of device-side and cloud-based processing. At the same time, Apple also announced a partnership with OpenAI to integrate its well-known language model ChatGPT into Siri. Although Apple has been committed to building its own language model, Federighi also admitted that there are other excellent large-scale language models on the market (LLM).
“These cutting-edge, ultra-large-scale language models have exciting capabilities that can bring more convenient experiences to users. We believe that integrating them into Apple products will make them more convenient for users to use,” Federighi said. GPT-4o is one of the best performing LLMs currently, which is why Apple chose to integrate it into Siri.
To protect user privacy, when "Apple Intelligence" cannot answer a user's question, Siri will ask for the user's consent before sending the request to ChatGPT for processing. Federighi revealed that Apple also plans to add support for more third-party language models in the future, allowing users to choose the most appropriate model according to their needs.
The Chinese market is one of Apple’s most important markets.Federighi said Apple is working to find ways to bring "Apple Intelligence" to the Chinese market. "While there is no specific timeline, this is certainly something we want to do.""Apple Intelligence" will initially only support US English, and it is not yet clear whether users in other regions will be able to use this feature.
IT Home noticed that during the interview, Federighi also talked about other topics related to "Apple Intelligence". For example, some user requests will be processed by Apple's own artificial intelligence server called "Private Cloud Compute". Apple has developed a method to encrypt and destroy data requests after processing them, thereby ensuring that user privacy information is not leaked.
While Federighi expects future chips to be able to run even larger language models, he also stressed that having online models will still be important to provide users with the latest information. "I'm not ruling it out," he said, "but even in that future, I think the device will still need to access external knowledge bases when processing user requests. So,Even in the future when processing can be done completely offline, interaction with external services will play an integral role.. "