Apple Intelligence is Apple’s interpretation of artificial intelligence, and it aims to revolutionise the way we interact with technology by integrating sophisticated machine learning and AI capabilities into commonplace devices.
Apple’s AI ecosystem is intended to improve user experiences and optimise operations throughout its product line by providing Siri with more conversational prose, automating revision and text summarisation across applications, and enabling lightning-fast image generation. The following is a comprehensive overview of Apple’s revolutionary new AI.
Apple Intelligence release date and compatibility
Apple Intelligence was initially scheduled for a formal release in September, which coincided with the release of iOS 18, iPadOS 18, and macOS Sequoia. Nevertheless, Apple subsequently decided to slightly postpone the release of Intelligence, as Bloomberg’s Mark Gurman reported. It is presently accessible to developers; however, Gurman believes that it is unlikely that Apple Intelligence will be made publicly available prior to the October 18.1 rollout.
The company has indicated that the AI features will be accessible on the iPhone 15 Pro and 15 Pro Max, as well as iPads and Macs with M1 or newer processors (and presumably the iPhone 16 handsets, as they will all be running iOS 18). This is the case, at least initially. Additionally, the features are exclusively accessible when the user language is configured to English at the time of launch.
What is the rationale behind the cutoff? Apple has maintained that the processes are excessively computationally intensive for older hardware, as they employ the more sophisticated neural engines, GPUs, and CPUs of the most recent processors.
In July, Apple’s Developer program granted users who operate an iPhone 15 Pro or iPhone 15 Pro Max access to an early version of Intelligence with the release of iOS 18.1 beta.
New AI features
The primary functions of Apple Intelligence will be to enhance Siri’s cognitive capabilities, provide writing assistance, and create and modify images, regardless of the device being used.
Apple Intelligence is intended to encompass the entire product portfolio. As a result, the iOS and iPadOS versions of Apple Intelligence will replicate nearly every feature that is present in the macOS version. This encompasses Siri’s enhancements, Image Playground, Memories in Photos, and Writing Tools.
Furthermore, the combination of iPadOS and Apple Pencil will grant access to additional capabilities. For instance, the Notes app’s Smart Script feature will automatically align and refine handwritten text in real time. The new Math Notes calculator will generate interactive graphs based on written equations in the user’s own handwriting and solve them automatically with a single stroke.
We at Digital Trends have already experimented with Apple Intelligence on the MacOS Sierra beta, but we have been somewhat underwhelmed by the digital agent’s performance thus far. For instance, the beta release only provides access to a small number of AI tools. Additionally, the tools that we did have access to, such as Siri, the writing assistant, and audio transcription, were unreliable and unstable. Granted, these are prototypes and will undoubtedly undergo additional refinement prior to public release. However, based on the evidence we have gathered thus far, Apple Intelligence’s anticipated October release may prove to be an overly ambitious deadline.
Writing Tools
Upon its release, iOS 18, iPadOS 18, and macOS Sequoia will incorporate the new Writing Tools feature. It has the capability to summarise text across Apple’s application ecosystem, including Mail, Notes, and Pages, and proofread the user’s writing, rewriting sections as needed. Third-party developers will have the ability to integrate Writing Tools into their own applications through API queries.
For instance, Apple Intelligence will present the user with concise summaries of the contents of their inbox in the Mail app, rather than the initial few lines of the email. Based on the contents of the message, Smart Reply will recommend responses and guarantee that the reply addresses all of the enquiries raised in the original email. Additionally, Priority Messages will elevate correspondence that is more pertinent and timely to the summit of the inbox.
Additionally, there will be substantial enhancements to the Notes application. Apple Intelligence will provide Notes with new audio transcription and summarisation capabilities, as well as an integrated calculator called Math Notes that resolves equations entered into the note’s body.
Image Playground
Users will have the ability to generate images in one of three artistic styles—Animation, Illustration, or Sketch—within seconds. The new Image Playground program will primarily manage the creation and editing of images. Image Playground will function as a standalone application, with a significant number of its features and functions integrated with other Apple applications, such as Messages.
Additionally, Apple Intelligence will be implemented in your camera roll. The Memories function in the Photos app was already capable of automatically identifying the most significant people, locations, and pets in a user’s life. It then curated this set of images into a coherent collection set to music. Memories is becoming even more impressive with the addition of Apple Intelligence.
The AI will choose the photographs and videos that most closely align with the user’s input prompt, such as “best friends road trip to LA 2024.” It will subsequently produce a narrative, which will consist of chapters that are based on the themes that the AI identifies in the selected images. The narrative will be assembled into a brief film. Users of Photos will also have access to Clean Up, a tool that is similar to Google’s Magic Eraser, and enhanced Search capabilities once Apple Intelligence completes the public beta.
Siri may be the most significant beneficiary of Apple Intelligence’s new capabilities. Apple’s digital assistant, which has endured for an extended period, will be further incorporated into the operating system, resulting in enhanced natural language processing and more conversational speech.
Additionally, the agent will be able to recall specifics from previous conversations, as Siri’s memory will be more enduring. Additionally, the user will be able to transition more seamlessly between spoken and written prompts.
Apple Intelligence privacy
Apple has prioritised privacy in the Apple Intelligence experience to prevent the costly and embarrassing data leaks that some of its competitors have experienced in recent months. The company has even invested in the development of its own private and secure AI compute cloud, Private Cloud Compute (PCC), to support intricate user enquiries.
Craig Federighi, Apple’s senior vice president of Software Engineering, stated at WWDC 2024 that the majority of Apple Intelligence’s routine operations will be conducted on-device, utilising the company’s most recent iterations of A17 and M-family processors. “It is cognisant of your personal data, but it does not collect it,” he continued.
Federighi continued, “Apple Intelligence evaluates whether a request can be executed on-device when it is submitted.” “If it requires a greater computational capacity, it can utilise Private Cloud Compute and transmit only the data that is pertinent to your task for processing on Apple silicon servers.” This should significantly decrease the likelihood of private user data being intercepted, compromised, spied upon, or otherwise snooped upon during transit between the device and PCC.
He clarified that Apple does not store or have access to your data. “It is exclusively utilised to fulfil your request, and, similar to your iPhone, independent experts can examine the code that operates on these servers to confirm this privacy commitment.”
Apple Intelligence’s ChatGPT partnership
Apple Intelligence will not be the sole cutting-edge generative AI that occupies your Apple devices this autumn. Apple and OpenAI executives announced at WWDC that the two companies are collaborating to integrate ChatGPT functionality (powered by GPT-4o) into Siri and Writing Tools. This partnership will include image and text generation. In the event that Siri’s integrated capabilities are insufficient to address the user’s inquiry, ChatGPT will intervene, similar to Apple Intelligence. However, ChatGPT will direct the request to OpenAI’s public compute cloud rather than the PCC.
When utilising ChatGPT’s capabilities, users will not be required to navigate away from the Siri interface; Siri will provide the response regardless of the query’s location. In order to guarantee a minimum level of privacy protection, the device will request the user’s consent prior to transmitting the request, as well as any documents or images that the user has attached.
Apple Intelligence was instructed on Google’s Tensor Processing Units.
Apple published a research paper in July that discloses the company’s decision to train critical components of the Apple Intelligence model using Google’s Tensor Processing Units (TPUs) rather than Nvidia’s highly sought-after GPU-based systems. According to the research team, the utilisation of TPUs enabled them to efficiently leverage the necessary computational power to train their extensive LLM, surpassing the energy efficiency of a standalone system.
This represents a substantial departure from the conventional business practices that are commonly observed in AI training. Currently, Nvidia holds an estimated 70% to 95% of the AI processor market. Therefore, it is exceedingly unusual for Apple to publicly disclose that it has chosen the product of its direct competitor. It may also serve as an indicator of forthcoming events. Nvidia’s market dominance is not sustainable indefinitely; hyperscalers of today are already transitioning to proprietary processor production. In addition to Google’s ongoing TPU initiatives, Amazon recently disclosed that it is developing its own chip line. This chip line is expected to surpass Nvidia’s current offerings by 50% while consuming half the power. In May, Microsoft disclosed that it would implement AMD’s AI processor line.