**Apple** has rolled out its Siri AI assistant, which it redesigned after nearly two years of delay, onto iPhones. This comes as the official release of iOS 27 is rolled out. However, with both usage caps and regional restrictions in place at the same time, the range of features users can actually enjoy has been significantly narrowed.
Key points
Siri AI is first provided in English beta on iPhone 15 Pro and above.
AI features have been added to built-in apps such as Photos and Safari, but cloud-based features are subject to a daily usage limit.
Early evaluations are mostly that “the app’s usefulness is high, but the model performance lags behind competing AI.”
iOS 27 Siri AI launch status
On September 14, Apple released iOS 27 for free updates to iPhone devices 11 and later. After going through developer and public betas throughout the summer, the official version has now arrived. Apple says it “speeds up app launch times by up to 30%.” However, most of the real “new features” can only be used on iPhone 15 Pro or later.
The new Siri AI is a fully redesigned assistant that uses the user’s messages, emails, and photos as context, answers questions based on what’s currently on screen, and pulls the latest information from the web. It kicked off first as an English beta along with a dedicated Siri app, with support for French, Japanese, Korean, Portuguese, and Spanish scheduled for October.
Siri AI, photo editing tools, Image Playground, and other server-based features all have daily usage caps set. Apple has signaled that it will expand usage limits through paid options afterward.
There are also big regional gaps. At first, Siri AI will not be available for EU iPhones, and in China, new AI features remain locked because the regulatory approval process hasn’t been completed yet.
The Photos app adds features like “Spatial Reframing,” which lets you rearrange the composition after taking a photo; “Extend,” which widens the photo edges; and “Clean Up,” an improved eraser with stronger performance. However, early tests have also led to feedback saying not to apply Reframing too aggressively.
Safari can automatically organize tabs by topic, indicate price drops, and generate custom expansion features based on only a brief explanation. It also adds a new slider that lets users directly adjust the intensity of the “Liquid Glass” visual effect.
Read together: XRP futures hit a 6-month high… Usage metrics soar, but prices stay flat
Federico Viticci on seeing Siri AI
In a review of iOS 27 released this week, Federico Viticci, a specialist editor at Apple, said that Siri AI’s baseline model performance falls behind competing products once it goes beyond simple knowledge Q&A. When it comes to deep research or analysis, he said it’s hard to replace OpenAI’s ChatGPT, and he also cited the lack of an “agent mode” (a mode that continues tasks autonomously) as a limitation.
Even so, Viticci diagnoses that Siri AI can do things inside iOS’s built-in apps that no third-party agent can really match. Based on that level alone, he expects it will become “a tool that is more than sufficient for millions of users.”
Apple trained five new “foundation models” using its own data, then refined performance using outputs generated by **Google**’s Gemini. However, Viticci emphasized that “there is no Gemini model actually running inside Apple software.”
This launch is the end of a long wait. At its 2024 developer conference, Apple promised a more personalized Siri with stronger contextual awareness, but as the release was delayed, it received criticism that it fell behind in the AI race. In the end, on June 8 at the WWDC keynote, Apple unveiled Siri AI again, opened a developer beta later that month, and launched a general beta program in July.
Next article: Iran sold oil, and Chinese companies received it in cryptocurrency… The U.S. Department of Justice targets $61 million