(Image credit: Getty Images) |
Assuming Apple introduces this year's iPhone software upgrade at WWDC 2024 on June 10, we won't be able to see iOS 18 for another six weeks. But if information on Apple's AI plans is true, the firm may have already made a very smart choice regarding the iOS update.
As you are undoubtedly well aware, Apple is aiming to include more AI-powered capabilities into its phones this year, and AI is expected to be a major emphasis of iOS 18. And it presents the business with an option: does it use the neural engine of the iPhone's system-on chip to ensure that everything operates locally or does it rely on the cloud to power those features?
According to reports, Apple has already decided to implement on-device AI for iOS 18. If accurate, that reveals a lot about the goals Apple has set for this year's upgrade.
Why On-Device AI Matters: A Crucial Advancement
The iPhone would have likely received additional functionality if cloud-based AI had been used. However, it would have come at the price of two aspects of the iPhone experience that Apple values highly: privacy and responsiveness. As Mark Gurman of Bloomberg puts it, in the case of on-device AI, "Apple's AI tools might not always be as strong and knowledgeable as they could be (the company could make up for this by partnering with Google and other AI providers), but the approach will expedite response times significantly." Additionally, Apple's ability to protect privacy will be easier."
"Apple’s AI tools may be a bit less powerful and knowledgeable in some cases (the company could fill in the gaps by teaming up with Google and other AI providers), but the approach will make response times far quicker. And it will be easier for Apple to maintain privacy."
— Mark Gurman, Bloomberg
Utilizing cloud servers expands the available processing power for certain functions. For instance, the Video Boost function on Google Pixel 8 Pros allows users to utilize the cloud to enhance the appearance of recorded video. However, compared to on-device translation, transcription, and summaries, which a mobile processor can finish quickly, that type of processing can take hours.
Regarding privacy, while businesses may implement safeguards to protect data sent to the cloud, the truth is that you have no control over data once it leaves your phone. It makes sense that Apple would take it into account with AI, as it has demonstrated sensitivity to it in other contexts.
Choosing on-device AI features might have one more drawback: it's unclear if all of the new capabilities Apple brings with iOS 18 will be compatible with all phones that can run the operating system. The next software upgrade, iOS 18, may be compatible with devices that date back to the 2018 iPhone XR, according to rumors. However, do iPhones from five or six years ago possess the same level of processing power as the iPhone 15 models that are now on the market or the iPhone 16 devices that will be introduced alongside iOS 18?
On the other hand, it's common for older phones to be left out of some capabilities that are added with an iOS update. With iOS 17, FaceTime now features on-screen effects that you can control with hand movements. However, in order to utilize this feature, your iPhone model must be at least 12 years old. Whether or not they are AI-powered, we anticipate that iOS will have functionality with comparable needs.
Anticipating On-Device AI: What's Ahead
(Image credit: Tom's Guide) |
What features will the iPhone get in iOS 18 if Apple's decision to reject cloud computing implies its AI skills won't be as strong? According to Gurman's reporting, AI in iOS 18 "plans to show how the technology can help people in their daily lives," as opposed to taking the form of chatbots.
A few AI models that Apple has previously made public offer further details about its intentions. It is believed that the majority of these models will be used in the development of Siri 2.0, an upgraded iteration of Apple's digital assistant that will be able to answer increasingly complicated questions.
- Ferret-UL is a multimodal large language model that can observe the screen of your phone and carry out tasks based on what it sees. For example, it can identify different sorts of icons or locate certain text.
- To improve multimodal model training, MM1 use synthetic data that includes text and visuals.
- Researchers and developers from outside the community may utilize OpenELM's collection of tiny language models to talk about the dangers, biases, and degrees of trustworthiness of the models.
It's unclear what those indicate in terms of specific iOS 18 features, but aside from the much-discussed Siri makeover, iOS 18 rumors have hinted at features like AI-generated playlists in Apple Music, Keynote slide creation, suggested text for Pages documents, and a Safari browsing assistant that might be able to summarize web page content.
iOS 18 AI: A Future Perspective
Of course, concrete information on AI-powered features is probably not going to be revealed until after Apple's iOS 18 debut at WWDC in June. Nevertheless, rumors about Apple favoring on-device features might help shape future expectations. Additionally, it appears that the focus will be on features that are responsive and capable of handling routine activities with grace.