Monday June 10, during its annual developer conference (WWDC), Apple announced ambitious software features powered by artificial intelligence (AI). Some will be added to Apple brand devices by the end of the year, others during 2025. The company is doing much better than catching up with the competition in terms of AI: the new features, grouped under the terms “Apple Intelligence”, constitute one of the most frank advances that iOS, iPadOS and MacOS software have known. But they will be reserved for the iPhone 15 Pro as well as computers and tablets running at least M1 chips, i.e. only a small part of the devices in circulation today.
A more resourceful Siri
If Siri’s developments live up to Apple’s announcements, it will indeed be a ” giant step “ for his assistant, to use the words used during his presentation. No need, to begin with, to be careful to use the right wording when speaking to Apple’s personal assistant: Siri should better understand our requests. It should also now be able to conduct complex searches inside an iPhone, iPad or MacBook, such as “show me photos of Stacy in New York with a pink coat.” The new Siri is also supposed to be able to control applications on an iPhone “with hundreds of different commands” : we can ask, for example, that the colors of the photo we have just found be brought to life.
With a single spoken command, Apple promises, Siri will summarize the notes taken during a meeting before sending them to a group of contacts by email. Or will look for our driving license number in the photo gallery to insert it into a form. It should be able to juggle between applications: according to Apple, to respond to a complex order like “can I push this appointment back an hour without risking missing my daughter’s dance performance?” », Siri will be able to find the time of the show in our emails, then consult our calendar, and find out about the road traffic situation, before providing an informed response. Finally, for information searches on the Internet, Siri will use Chat GPT-4o, OpenAI’s conversational agent, at no additional cost.
It remains to be seen whether this personal assistant with renewed ambitions will be able to carry out the vast majority of tasks entrusted to him, or whether he will fail too often for us to trust him. In spirit, the new Siri seems in any case much more ambitious than the new assistant whose contours Google drew a few weeks ago during its own developers conference.
The new Siri, at least as it is intended by Apple, is intended to be more than a voice assistant, but simply an assistant, capable of assisting us in all circumstances. In fact, this new version will no longer be controlled by voice alone: it will also be possible to give it written instructions.
Read also | Google will slip its Gemini AI into Android and its search engine
Add to your selections
Tools for writing
When writing a note, an e-mail, a text message, it will be possible on the machines concerned to bring up a writing assistance tool to rewrite the message in a different tone (friendly, professional, concise) or ask ChatGPT to write a draft from scratch, based on a short instruction.
The new features announced by Apple will also be able to summarize incoming emails and respond to them intelligently. To better prepare the answer, questions will be submitted to the user, in the form of a quick MCQ, which will serve as a basis for writing the message. One of the most surprising features: the one which offers to only display the program e-mails and notifications deemed priority by the program, and even to “reduce interruptions” by only letting important alerts go through. However, are we prepared to take the risk of entrusting an AI with the power to wrongly block an alert that could prove important? Only an in-depth test will allow us to test its reliability.
Tools for creating images
Several AI-powered tools should make it possible, on the machines concerned, to generate entirely new images based on a simple textual instruction and to insert the face of a loved one – or one’s own. These visuals may resemble a drawing, an image, or an animated image. Apple will also offer to create fully personalized emojis, in order to illustrate a message for which no existing emoticon is naturally required. In addition, it will be possible, from existing videos, to create a short film by typing a simple instruction: the program will be responsible for editing the videos. With a talent that is currently very difficult to determine.
Please note: we will now be able to record voice calls on iPhone and obtain an automatic transcription.
A framework for privacy
To carry out these tasks, the device will often use its own artificial intelligence programs, running directly inside its chips, in order to preserve the privacy of its users. Apple does not specify what impact these calculations will have on battery life.
The world
Special offer for students and teachers
Access all our content unlimitedly from €9.99/month instead of €11.99.
Subscribe
For certain requests, however, the device will have to contact an Apple data center via an Internet connection. Apple promises that only the most strictly necessary personal information will be sent to these remote computers. These should theoretically only be used to process the request, before being immediately deleted. Clearly: they are not intended to be used to train Apple and OpenAI models.
Improvements for iOS and MacOS
In addition to the announcements on AI, Apple has announced dozens of software developments intended for its devices. Two among others: the software that equips iPhones will now be able to lock an application, so that it is only unlocked when you present your face to the FaceID system or enter the mobile password. Finally, on Mac, we will soon be able to display the iPhone screen on the computer screen, and manipulate all its applications with the mouse. For, for example, dragging and dropping files from the computer to the smartphone.
