Apple’s otherworldly flying headquarters in Cupertino, California, seemed like the right place this week for a bold and futuristic overhaul of the company’s most prized products. As iPhone sales decline and competition increases with the development of tools like ChatGPT, Apple unveiled its own vision for generative AI at its Worldwide Developer Conference (WWDC).
Apple has been seen as a generative AI laggard lately. The WWDC offer did not convince some critics, who described the WWDC announcements as downright boring. But focusing on combining existing apps and operating system features with what the company calls “Apple intelligence,” the most essential takeaway is that generative AI is a feature, not a product in itself.
The dazzling capabilities demonstrated by ChatGPT have inspired some startups to try to invent completely dedicated AI hardware – such as Rabbit R1 and Humanitarian AI pin—as a way to operate generative artificial intelligence. Unfortunately, using these gadgets in practice is disappointing and frustrating. In contrast, vertical integration of Apple’s generative AI across so many products and various software seems much more likely to be the direction AI is heading.
Instead of creating a standalone device or solution, Apple focused on how generative AI can improve applications and operating system features in diminutive but significant ways. Early adopters certainly started using generative AI programs like ChatGPT to support rephrase emails, summarize documents, and generate images, but that usually meant opening another browser window or application, cutting and pasting, and trying to understand the chatbot’s sometimes frantic gibberish . To be truly useful, generative AI will need to make its way into the technology we already operate, in a way we can better understand and trust.
Following its WWDC keynote, Apple gave WIRED a demo of what it calls Apple Intelligence, a catchy name to describe artificial intelligence that runs across several applications. The capabilities don’t push the boundaries of generative AI, but they are thoughtfully integrated and perhaps even narrow in ways that encourage users to trust them more.
A feature called Writing Tools will let iOS and macOS users rewrite or summarize text, while Image Playground will turn sketches and text prompts into stylized illustrations. The company is modern Genmoji toolwhich uses generative AI to create modern emojis based on text prompts, could prove to be a surprisingly popular integration considering how often emojis are thrown at each other.
Apple is also giving Siri a much-needed generative AI update that helps the assistant better understand speech, including pauses and corrections, recall previous conversations for better context awareness, and operate data stored in apps on the device to be more useful. Apple said Siri will operate App Intents, a framework for developers to use to perform actions including opening and operating the application. For example, when asked “show photos of my cat chasing a toy”, the language model will parse the command and then operate the structure to access the Photos.
Apple’s generational AI will mostly run locally on Apple devices, although the company has developed a technique called Private Cloud Compute that allows it to securely query the cloud when needed. Running AI on a device means it will be less capable than the latest cloud-based chatbot. However, this may be a feature rather than a bug, as it also means that a program like Siri is less likely to over-expand and break. Apple quite cleverly forwards the most complex queries to OpenAI’s ChatGPT, with the user’s consent.