By unveiling its version of AI, Apple put forward a different message from the competition: Apple Intelligence wants to be useful and personal rather than artificial and “wow”. A popularization bet as daring as it is risky.
It is now an actor to be officially taken into account. Apple has launched into the battle of artificial intelligence, but in its own way. After the first shot initiated by Microsoft, Open AI and Google, Tim Cook and his team let the storm pass, waiting for the ground to clear the superficial and the broken pots to come and present their solution. It’s called Apple Intelligence and, just in its name, already announces the vision that will drive it.
Because Apple did not want to make a big splash or “wow” announcement for its grand induction during the opening keynote of the WWDC developers conference at Apple Park. No functions that leave you speechless, no revolutionary new features that undermine the competition. That’s not how the owner of the place presented it anyway. Apple Intelligence aims to be “powerful, secure, useful and above all personal”. Four words to sum it all up and a roadmap deployed much more cleverly than one might think.
No “wow” effect sought
It is also amusing to note that the notion of “artificial” has completely disappeared. If we heard “artificial intelligence” once during the 45 minutes of presentation devoted to it, that’s it. As always at Apple: it’s others who say it — journalists and observers alike, not them. Behind the scenes, we nod and smile when we ask if it’s because artificiality has no place at Apple.
The latter also replaced it with… Apple in the term. It’s the intelligence of your Apple product at your service. Just what is needed to replace the now sacrosanct letters of the term with its own (AI). A well-executed communication plan which says a lot, especially about the desire to popularize everything to be understandable to as many people as possible.
Because this AI, whose generative language models have all been designed in-house, wants to do everything so that the use of the iPhone, iPad or Mac which will be equipped with it is as natural and as obvious as possible, anchored in everyday life. And Apple repeats that it will be optimized to understand us, support us in writing/correcting messages such as photo editing or transcription, understanding what we need. In short, standard uses. Your questions about life, cooking, the child’s lesson, will be for others.
From what Apple Intelligence revealed, nothing was new in the uses. Apple has done Apple by making this understandable to anyone, integrated directly so that usage is just a click away. It is already a feat in itself to be able to popularize AI to the point of losing what many have made a selling point: a language that the philistines cannot understand and which gives the impression of floating above the masses ready to join them, therefore to be important.
Apple wants simple and useful AI. It must be in language. The brand, although accustomed to praising the calculation values of its chips, has never spoken of tokens by hundreds of thousands to give the impression of superpower with incredible results. In fact, we know that more does not mean more efficient in AI.
Instead, Apple has updated Siri, now powered by generative AI, to better understand the user and interact as naturally as possible by taking much more into account the data stored on their device.
Data security, the punchline argument
Because where Apple also wants to stand out from other generative AI is in its approach to data confidentiality and security. And in this, Apple could not risk falling in love with OpenAI a little too much without safeguards. The company has also warned the father of ChatGPT on this subject, but also the users.
If the iPhone user is willing to share their data with ChatGPT to get more personal answers, they will do so with full knowledge and validation. But Open AI must ensure the security of the data and not keep it for its personal training use.
This will be the case for the other specialized generative AI bricks which will be added over time. Apple is studying the issue to gradually add language models on specific points (health, leisure, understanding, information processing, etc.). This is also why Google Gemini has not been ruled out and should very soon be included in the adventure.
For the moment, it is difficult to know when Apple Intelligence will reach its true cruising speed; probably in 2025. Apple’s AI will only be accessible to developers in the summer and therefore not available in the beta phase of iOS 18, iPadOS 18 or macOS Sequoia deployed in July.
We will have to wait for the final versions in the fall to get a first glimpse – in English and possibly only in the United States at first, with already the preventive message: “it will not yet be stable and we will still have to wait for the final deployment.” What will we have to eat? Will the iPhone become an even smarter weapon? Nobody knows. This is a vision that Apple has given, its roadmap. And the road is still long…
