Last year, Apple launched a similar plan called private cloud computing with its Apple Intelligence AI platform. Users can open the service in Apple’s end-to-end encrypted communications app to generate a message digest and write a “smart reply” message on iPhone and Mac.
However, looking at private cloud computing and private processing is like comparisons, Apple(S) and Orange. Apple’s private cloud computing foundation, all Apple intelligence can be applied everywhere. Private processing, on the other hand, is built specifically for WhatsApp and is not a broader Meta AI feature. Apple Intelligence is also designed to do as much AI processing as possible and to send only requests to private cloud computing infrastructure if necessary. Because such “device” or “local” processing requires powerful hardware, Apple only designed Apple Intelligence to run on its recent generations of mobile hardware. Older iPhones and iPads will never support Apple Intelligence.
Apple is a manufacturer of high-end smartphones and other hardware, while Meta is a software company with about 3 billion users with various types of smartphones, including old and low-end devices. Rohlf and Colin Clemmons, one of the chief engineers in private machining, said it was not feasible to design AI capabilities for WhatsApp, which can run locally within the range of devices that WhatsApp serves. Instead, WhatsApp focuses on designing private processing to make attackers as helpless as possible.
“Design is one of the risks minimization,” Clemens said. “We want to minimize the value that damages the system.”
However, the entire effort raises a more fundamental question of why secure communication platforms like WhatsApp need to provide AI capabilities at all. However, the Yuan people firmly believe that users expect these features at this point and will acquire them wherever they are needed.
“Many people want to use AI tools to help them when sending messages,” WhatsApp Head Will Cathcart told Wired in an email. “We think it’s important to build a private approach because people don’t have to switch to a private platform to have the functionality they need.”
“Any end-to-end encryption system that uses partial AI reasoning will be more risky than a pure end-to-end system,” said Johns Hopkins cryptographer. “I believe WhatsApp says they have designed this to be as safe as possible, and I believe they say they can’t read your text.” But I think there are risks here, too. More private data will be disconnected from the device, and the machines that process this data will be targeted by hackers and nation-state rivals. ”
WhatsApp also says that in addition to basic AI features, such as text summary and writing suggestions, private processing will hopefully expand to more complex and involved AI capabilities in the future, involving processing and potentially stored more data.
As Greene said, “All this will make a private machining computer a very big goal considering all the crazy things people use for security messengers.”
