Tim Cook at Apple WWDC 2025.
Jason Hiner/ZDNET
One of my editors once told me that a cynic will always sound more rational, but that doesn’t mean they’re right. French statesman Francois Guizot basically said the same thing from the other POV: “The world belongs to optimists. Pessimists are only spectators.”
Apple’s AI strategy has a lot of cynics and pessimists — and they sound pretty rational based on what we’ve seen over the past year. The long-awaited revamp of Siri has never materialized and is reportedly delayed to 2026. Apple’s vision of your own custom AI-powered assistant with your “personal context” has been little more than a hopeful vision. And the AI features that have arrived — such as summaries of a string of text messages — have been both game-changing and, at times, super frustrating to use.
Also: Apple’s de-chatbot-ification of AI is nearly complete
Still, the generative AI revolution that is sweeping the tech industry is in desperate need of many of the qualities that Apple usually exemplifies in its approach to products. I’m talking about qualities like polish, attention to detail, user trust, and patience.
While generative AI is racing forward at a breakneck pace, it continues to feel half-baked and overhyped much of the time. That’s likely why ZDNET/Aberdeen research showed that only 8% of the broader public say they’re willing to pay for AI features on their devices. Even more challenging, 69% said they would stop using a product or consider stopping if it had AI features that they couldn’t turn off.
The AI enthusiasm gap
In other words, our research shows a massive gap between the next-big-thing narrative that companies like OpenAI, Google, Anthropic, and others are telling us about the AI revolution and the lack of enthusiasm from the people who are learning about these new AI features and trying to use them in the real world.
CNET survey: Just 11% of people upgrade their phone for AI features. Here’s what they want instead
Don’t get me wrong, there’s a lot to be excited about. If you’ve used one of these generative AI tools to summarize a 500-page document for you, help you write a recommendation letter, or translate a document from a relatively obscure language into English, then you’ve gotten a taste of how useful they can be in the right circumstances. And if you’re a coder or a programmer, generative AI is a life-changer.
But if you’ve used chatbots for much time at all and started double-checking their accuracy, then you’ve likely also come across the fact that products like ChatGPT regularly hallucinate, make stuff up, and simply get things wrong. In short, they can’t be fully trusted, and that limits a lot of their usefulness because you often have to double-check their work.
(Disclosure: Ziff Davis, ZDNET’s parent company, filed an April 2025 lawsuit against OpenAI, alleging it infringed Ziff Davis copyrights in training and operating its AI systems.)
Also: Apple’s Goldilocks approach to AI at WWDC is a winner. Here’s why
Last year at WWDC 2024, Apple spent a lot of time trying to convince us that it’s an AI leader, that it’s been working on AI for years, and that it had exciting new AI breakthroughs on the way that only Apple could deliver because of its broad ecosystem of devices and platforms and its strong privacy stance.
It was a bold move since Apple rarely talks about future products and features until they are nearly fully baked. In retrospect, last year’s AI blitz at WWDC was unnecessary.
Apple didn’t need to enter the AI features arms race, in which OpenAI, Google, and a constellation of tech giants and cash-rich startups are playing a relentless game of one-upmanship.
That’s not where Apple can compete, and it’s not where Apple can make the biggest impact in tech’s next big wave of innovation.
Apple’s de-chatbot-ification strategy
Apple simply needs to run the Apple playbook. After WWDC 2025, it’s clear that Apple focused on integrating AI feature-by-feature into its existing software. It’s going to take the best features of LLMs and integrate them directly into its software, such as language translation in the Phone and Messages apps. There are a lot of existing features on the iPhone that could be enhanced, streamlined, and made more powerful — all outside of the chatbot.
When it comes to Siri, Apple has likely realized that it needs to start over from the ground up. The codebase for Siri goes back decades to old government projects in DARPA, and my understanding from people with knowledge of the code is that retrofitting it for today’s uses has created a classic innovator’s dilemma that has hindered Siri for a decade and kept it from keeping up with Amazon Alexa, Google Assistant, and now ChatGPT.
Also: Only 8% of Americans would pay extra for AI, according to ZDNET-Aberdeen research
That doesn’t necessarily mean Apple needs to acquire its way into competition with LLMs and the latest chatbots. Anthropic’s Claude regularly gets mentioned as a good fit for Apple because of Anthropic’s focus on privacy and AI safety. However, Anthropic was valued at $61.5 billion in March 2025, so an acquisition would cost over $100 billion. That makes it very unlikely Apple or anyone else is buying Anthropic.
Apple has AI talent, and it has time. The generative AI revolution is still just beginning to take flight. Even if it took two years to rebuild the foundation of Siri for the age of LLMs, it would be worth it — and lots of users will still be at the very beginning of their journey with AI in 2027.
Meanwhile, I expect that we’ll continue to see Apple use LLMs to enhance individual features in the iPhone and throughout its various devices and platforms — but do it in the Apple way. Wait until the features are fully baked before showing us why we should be excited. That’s what we got at WWDC 2025 as Apple returned to being more Apple-like in its announcements about future products and features.
We also shouldn’t forget where the best generative AI chatbots, products, and features from all the biggest innovators are running today and will be running for years to come — on laptops, smartphones, and tablets, many of them made by Apple.
There are over 50,000 AI startups, according to Crunchbase and over $27 billion is currently being invested in AI, according to the New York Times. All of these new apps and tools that are being created will give people even more reasons to use their favorite devices and Apple has building chips with neural engines.
PCMag: Apple’s WWDC Was All Glass, No Vision: What’s on Tap to Counter OpenAI, Google?
And while OpenAI and Jony Ive have made a lot of noise about creating their own AI hardware device, even they’ve admitted that it’s not going to take the place of your phone or your laptop for running the latest AI tools.
So, yes, Apple remains very well-positioned to play a leading role in the AI innovation wave that’s in front of us. The AI world needs Apple’s discipline, focus on user experience, and patience to play the long game. And let’s be honest — it’s going to need a lot of Apple’s most powerful devices as well, no matter whose software is running on them.
