Apple introduces new digital assistant Siri AI

Apple is late finally introduced a modern version of its digital assistant, called Siri AI. The assistant is Apple’s equivalent of Gemini for Android. The new Siri is also based on Gemini models through an agreement with Google.

According to Apple, the new Siri is far more capable than the old Siri. The assistant must understand personal context, have knowledge of the world, be able to retrieve current information from the web, know what is shown on the screen and be able to answer almost any question regardless of topic. Just like Gemini – at least in some parts of the world – Siri AI can also pick up relevant information from users’ data such as messages, emails and images.

The “personal context” can be extended to third-party apps as long as the developer offers Spotlight integration. For example, users should be able to ask Siri AI to find a restaurant a friend previously recommended, show the confirmation number for an old hotel reservation, or retrieve optional photographs from the archive.

A profoundly more capable and conversational assistant with personal context understanding, broad world knowledge, and onscreen awareness, Siri AI can help users find what they need in the moment, from answering questions from the web on virtually any topic, to surfacing relevant information from a user’s personal messages, emails, photos, and more.

Since Siri, like Gemini, has access to the screen content, it is possible to start the assistant and ask about things displayed in the interface. Apple also mentions do-it-yourself features that have become standard, such as writing email drafts and notes and editing photos.

Another big news is that Siri got a separate app, for example to resume conversations like in Gemini, Claude and Chat GPT. In IOS 27, the assistant can be launched through the voice command “Hey Siri”, through the power button or by swiping down from the “dynamic island” at the top of the screen (in Android, Gemini can be launched by swiping up from the lower right corner of the interface).

On select unspecified models capable of running the most advanced local models, Siri AI will offer more expressive voices, as well as significantly better dictation. Users can control both how fast Siri should speak and how expressive the voice should be.

The integrity

Siri AI leverages both local models and the cloud. A solution called Private Cloud Computing shall improve privacy by ensuring that users’ personal data is neither stored nor made available to Apple or anyone else. Apple invites outside experts to verify the privacy promise.

Privacy shall specifically refer to users’ “personal data”. So Apple does not explicitly say that no conversations will be reviewed by humans. A major disadvantage regarding the privacy of digital assistants for private individuals – be it Gemini, Chat GPT or Claude – is that everything that is said can be read by people after anonymization (even if the risk is small). We will return if it is confirmed that no one else can read what is said to Siri AI.

Availability

Siri AI will not only be released for iOS 27 but for all Apple operating systems. The assistant will be beta tested later this summer and it will debut in the 27 versions of the platforms in the fall.

New Siri for IOS 27 and Ipad OS 27 will be delayed in the EU, which Apple makes a big deal about in a separate announcement where the company blames regulation on digital markets (DMA). The Union has stricter privacy and competition laws than many other parts of the world.

However, the assistant should be released in Mac OS 27, Vision OS 27 and Watch OS 27 without delay in Europe as well.