Apple iOS 18 will offer on-device privacy instead of server-side privacy

A fresh report on Sunday repeats the rumor that Apple’s focus in iOS 18’s AI efforts will center on privacy, handling processing directly on the iPhone without connecting to cloud services.

In recent months, there’s been a buzz about Apple’s ventures into Artificial Intelligence and the features it plans to roll out later this year with iOS 18 and macOS 15. Many sources have chimed in, all suggesting that Apple will be bringing AI-related improvements.

MORE: Apple “Let Loose” Event Slated For May 7: New iPads Expected

It seems like the first set of features won’t require an internet connection. According to Bloomberg’s Mark Gurman, in his weekly PowerOn newsletter on Sunday, he mentioned that the initial batch of AI-related features that Apple intends to introduce with iOS 18 “will operate completely on the device.” In simpler terms, these AI features will be able to work even without internet access or any cloud-based processing.

According to sources familiar with the matter who spoke to AppleInsider, it seems like the report’s claims are spot-on. Apple is developing its own large language model (LLM), internally dubbed “Ajax.” Although the fancier features might need an internet connection down the line, the essential text analysis and response generation functions are expected to work offline.

In terms of specific apps, expect upgrades in Messages, Safari, Spotlight Search, and Siri. Apple’s been tinkering with on-device text-based response generation for some time now, so chances are high that it’ll be one of the first features teased or rolled out. However, for the fancier AI stuff, like the advanced features and improvements, you’ll still need to be online. Rumor has it that Apple mulled over using Google and OpenAI’s AI tech instead of building its own online LLM for iOS 18.

READ: VisionOS features that Apple should bring to iOS 18

The big surge in AI projects lately is probably what’s got Apple eyeing artificial intelligence. AI tools have become more common among regular folks, and they’re causing some headaches in the legal and education fields. Moving AI stuff onto devices could help squash some of the problems linked with server-based AI tools. Like, for instance, those tools sometimes make up info out of thin air, a phenomenon known as hallucinating.

The hallucination issue has actually escalated because AI models are feeding off content created by other AI models. With its new “Ajax” LLM, Apple probably aims to go head-to-head with the existing tools out there. By ditching the need for cloud processing and stepping up text generation quality, they could really outshine their AI rivals. We’ll get the full scoop on Apple’s AI game plan during WWDC, kicking off on June 10th.

While iOS 18 might not roll out with a chatbot like ChatGPT or Copilot, it seems like Apple’s focusing on adding generative AI skills to their own apps. The goal is to ramp up efficiency and features across a bunch of Apple services, from Siri and the Notes app to Spotlight search, Safari, Apple Music, and more.