Apple’s Generative AI: What It Means for Your iPhone
Apple, well-known for its history of innovative products, including the iPhone, iPad, Mac, and Apple Watch, is now taking a significant step into the world of technology that could reshape your iPhone experience. This pioneering advancement is none other than generative AI, positioned to alter the way you interact with your device.
Generative AI, a subset of artificial intelligence, possesses the remarkable capability to create fresh content encompassing text, images, audio, and video, all derived from provided data or prompts. To illustrate, when you instruct Siri to craft a love-themed poem, generative AI can cleverly compose a poem customized to your specific request.
The methodology behind generative AI involves the utilization of deep learning models that glean insights from copious datasets, spanning text, images, audio, or video. It subsequently employs various techniques to generate novel content mirroring the characteristics of the input data. Notable generative AI models in this realm include GANs (Generative Adversarial Networks), VAEs (Variational Autoencoders), and Transformers. GANs have two neural networks: one that makes fake content and one that checks if the content is real or not. They compete until the first one can trick the second one. GANs are good for making and changing images, like making realistic faces, changing hairstyles, or transferring styles.
VAEs have one neural network that compresses the data into a smaller space and then makes new content from that space. VAEs are good for making and fixing images, like making anime characters, restoring old photos, or filling in missing parts.
Transformers have one neural network that can process data in a sequence, like text, audio, or video. They use attention to learn how different parts of the data relate to each other. Transformers are good for making and understanding language, like making chatbots, writing essays, or summarizing articles. They can also make and mix audio and video, like making speech, music, or movies.
According to Jeff Pu, an analyst who follows Apple closely, Apple is planning to use generative AI on the iPhone and iPad in late 2024 or early 2025. This means that you will be able to use generative AI features on your device, like improving Siri’s language skills, automating complex tasks with Shortcuts, and boosting your creativity with various apps.
Apple will use both cloud-based AI and edge AI, which means that some of the data processing will happen on your device, and some will happen online. This way, Apple can keep your privacy and security safe.
Apple has many generative AI projects going on, like a chatbot called Apple GPT-3, a visual intelligence team that can make images, videos, and 3D scenes, and a multimodal AI team that can handle text, images, and videos.
Apple will probably use generative AI in Siri, its voice assistant to make it better at understanding and helping you with what you want. This feature might come out as soon as iOS 18 next year.
Apple might also use generative AI in different apps to help you create and work better, like Xcode, GarageBand, Logic, iMovie, Final Cut and Fitness+. For instance, generative AI could help you write code faster and more accurately in Xcode, or help you edit audio and video with just your voice or gestures in GarageBand or iMovie.
Apple is not alone in the generative AI field. Other companies like OpenAI, Google and Microsoft have also made their own chatbots and large language models. But Apple is careful and thoughtful about how to use generative AI in a way that avoids problems like bias, hallucination or misuse.
Generative AI is a fascinating and captivating technology with the ability to create fresh content from scratch. Apple has plans to integrate generative AI into the iPhone and iPad by late 2024 or early 2025, potentially transforming your device usage and unlocking your creative potential. Apple’s foray into generative AI represents a significant advancement for your iPhone, and it’s an opportunity you won’t want to overlook. Apple has also bought several AI start-ups in the past few years to improve its generative AI skills, like Voysis, Xnor.ai, Spectral Edge, and Lattice Data. These start-ups could help Apple make better voice, image, and data processing and generation technologies. Apple is spending a lot of money and time on generative AI research and development, and it will show us what it can do soon.