Apple might appear to have entered the generative AI field later than some of its competitors such as Google, OpenAI, Microsoft, and Anthropic. However, given Apple’s longstanding strategy of focusing on delivering superior products instead of being the first to market (“not first, but best”), this delay in investment could be seen as a strategic choice rather than a lack of commitment.
Last year, Apple unveiled its AI plan, dubbed “Apple Intelligence,” aiming to significantly enhance Siri by incorporating advanced features. This transformation would go beyond making Siri a basic digital assistant, enabling it to match the capabilities of ChatGPT and Microsoft’s Copilot on an equal footing.
Reflecting on the past year, it appears that Apple’s grand promises regarding their AI capabilities have yet to fully materialize. Even before the official unveiling of their AI suite across various tech platforms, there were whispers of apprehension among insiders. They expressed doubts about the maturity level of Apple’s AI technology, suggesting it might be trailing two years behind OpenAI’s ChatGPT in terms of development.
As a devoted fan, I can’t help but acknowledge that Satya Nadella, Microsoft’s CEO, stated that OpenAI had an uncontested two-year period to develop ChatGPT. In contrast, Apple has primarily rolled out minor AI features like Writing Tools and Image Playground, which some users have humorously dubbed “vaporware.
Previously this year, the company disclosed that the release of Apple Intelligence was being pushed back to 2026 without providing an explanation. This postponement of their AI plans appears to have caused issues for the makers of iPhones, as they now find themselves in court battling a lawsuit claiming they used “misleading” Apple Intelligence advertisements to boost sales of the iPhone 16.
Apple Intelligence had two iterations yet none shipped
I’ve been keeping an eye on the buzz surrounding Apple’s postponed rollout of Apple Intelligence. Recently, I had a chance to sit down with Apple’s Software Engineering Senior Vice President, Craig Federighi, and Global VP of Marketing, Greg Joswiak, to delve into their strategic approach towards AI.
It’s worth mentioning that Apple conducted its annual developer gathering, WWDC 2025, and revealed a variety of updates, including iOS 26 featuring Liquid Glass, a user interface resembling Microsoft’s Aero UI from Vista in Windows. Interestingly, Apple Intelligence was not mentioned at the event, hinting that Apple may unveil it separately at a later time.
During conversations with media outlets, Federighi offered additional insights about the postponed rollout of Apple’s project called Apple Intelligence.
As a researcher, I can share that during the development of our innovative feature, we identified two distinct phases or versions of its ultimate architecture. By the time we were nearing the conference, we had successfully implemented the first version and held a strong belief in our ability to deliver it, with high confidence at the time.
Initially, I anticipated unveiling our plans by December, though I admittedly hoped for springtime at the latest. However, when it came down to it, we decided to reveal this at WWDC because we recognized a strong global interest in understanding Apple’s perspective on the future of Apple intelligence.
The top executive disclosed that Apple is concurrently developing two designs for the Siri infrastructure. He mentioned that the second design, referred to as V2, encompasses a more comprehensive end-to-end architecture aimed at achieving a complete set of capabilities desired for Siri, with an emphasis on a deeper structure.
In simpler terms, he revealed that the demonstration videos presented at Apple’s WWDC 2024 were intended for the first version of the Siri architecture. Apple has been putting more focus and resources into enhancing Apple Intelligence and refining Siri, as indicated by their intensified efforts.
Nevertheless, Apple soon discovered that the V1 architecture had significant drawbacks, as it failed to meet the high-quality standards that they understood their customers demanded and desired.
It became apparent that the V1 structure, despite our best efforts to optimize it further by investing more time, wouldn’t measure up to our customers’ expectations or Apple’s quality standards. Consequently, we decided to transition to the V2 architecture.
Apple’s Senior Vice President of Software Engineering, Craig Federighi
Apple admitted during the springtime that they wouldn’t meet their initial deadline for releasing Apple Intelligence and the upgraded Siri as planned, choosing instead to make this information public.
Currently, Apple is directing its attention towards the development of the V2 Siri infrastructure. As revealed by Federighi, the company has taken lessons from past errors with the V1 architecture, choosing not to disclose any specific launch dates for Apple Intelligence or the revamped Siri at this time.
The executive made it clear that they would only talk about the progress of their AI strategy once the V2 system was consistently producing results within the company, not just offering video presentations as proof.
Craig Federighi playfully hinted at showcasing the V2 architecture Apple has developed, but chose not to, as he mentioned that it’s functional internally. However, he added that Apple isn’t quite there yet in terms of delivering quality that matches his standards for a truly exceptional Apple feature.
The leader implied that the tech titan won’t disclose the release date for its AI plan until they are fully prepared to implement it. Additionally, he offered additional insights regarding the Version 2 structure of Apple’s technology that they have been developing.
Initially, the V1 architecture encompassed just half of what V2 had to offer. Now, we’re expanding it comprehensively, aiming to create a complete architecture that spans the entirety of the Siri user experience.
Over the past while, I’ve been focusing on expanding and refining what we started for Version 1. This comprehensive overhaul is resulting in a more unified end-to-end structure, significantly improving the quality and capabilities of our work.
According to the executives, when they embarked on Apple Intelligence, it was never their intention solely to create a chatbot for the users. Instead, as emphasized by Federighi, “we made it clear from the beginning that this wasn’t just about building a chatbot.
Rather than separating tasks into different experiences, Apple strives to infuse Apple Intelligence throughout their technology ecosystem, making it seamless for users to accomplish tasks without having to switch between various interfaces.
Read More
2025-06-11 14:09