Apple has defended its decision to leave most iPhone owners behind when it pushes ahead with newly announced AI-powered features, set to roll out later this year only on future iPhones and premium models from the most recent generation.
AI boss John Giannandrea said this move had purely technical reasons, explaining in a live interview on The Talk Show podcast on Tuesday, that the AI software required more powerful hardware.
The AI models could theoretically also run on older phones, he admitted, but argued they would be too slow to be useful.
Apple announced an AI offensive earlier in the week with a wide array of “Apple Intelligence” features for iPhone users, notably editing help in emails, individual emojis and an upgraded Siri that harnsesses the smarts of ChatGPT.
However Apple is only bringing the features to future iPhones and the more expensive models of the iPhone 15 from late last year.
There is more leeway with Mac computers, and any of the in-house M-series chips, which Apple has been using to replace Intel processors since 2020, will suffice.
The AI features will also run on iPads with M chip systems. In addition to the chip, the RAM memory capacity also plays a role, software boss Craig Federighi said.
Apple, marketing itself as a privacy-focussed company, says it has sought to do as much computational work as possible on the devices for data protection reasons. This also increases the demands on their performance, however.
At the same time, the company has developed a procedure with which the tasks are forwarded to Apple’s servers in encrypted form. The data should then disappear completely from the cloud. “We don’t utilize user data to train our models,” Giannandrea said.
Arriving in updates like iOS 18 in the second half of 2024, the new features will also be able to generate images from text prompts. Unlike many other services, however, Apple AI does not generate artificial photos out of concerns users might create deepfakes.
Source Agencies
