Using Deep Learning, and Speech-to-Text transcription, Apple has introduced a refresh of the Voicemail service that many people will now find more useful than before. It takes advantage of the new M3 chip to convert the audio in your voicemail to written text in real-time. The conversion process will happen locally on your device, ensuring the privacy of your messages as no data is shared or stored on the cloud.
The iOS keyboard will now use a real-time transformer model to predict your next word in real-time. This is possible using the updated M3 silicon chip on the iPhone. By making use of the real-time transformer model, the keyboard can accurately process and predict text corrections with each keystroke. Furthermore, the keyboard adapts to your commonly used words and phrases which allows it to generate inline suggestions that can be added to your sentence by simply tapping on the spacebar. This adaptive learning mechanism aims to enhance typing speed and accuracy, resulting in a seamless and efficient user experience.
A new app is coming to iOS 17. Journal will utilize on-device machine learning to provide personalized suggestions tailored to your preferences. It will pull insights from various data sources on your iPhone, such as workout logs, photos, and location information to enhance your journaling experience and suggest specific formats.
You will have complete control over the suggestions generated by the Journal app. It will integrate seamlessly with other applications and services through the suggestions API, so if you're a developer you'll be able to use the provided API to expose data from your app to the Journal app. Apple also outlined that all data is protected with end-to-end encryption, ensuring that users’ information remains safe at all times.