There have been rumors for a while that Apple is developing its own AI product like ChatGPT and it may be released in the future. So, here’s everything we know about AppleGPT so far!
AppleGPT Coming Soon?
AppleGPT is currently an in-house generative AI chatbot used by Apple engineers. It is rumored to be released via Siri, Apple’s voice assistant. The company may give Siri a huge upgrade with such an integration.
The chatbot functions might be similar to Bing AI, ChatGPT, and Gemini. While engineers prototype features and use them to summarise material and respond to inquiries in their daily workflow, it is being trained on a sizable language model.
The upgrade can make Apple a formidable opponent for ChatGPT and Gemini in the upcoming days as it holds vast potential to be a leading giant in the AI race.
Google Jax, the machine learning framework used by Google, serves as the foundation for Apple’s AI framework, Ajax. Apple developed its Large Language Model (LLM), which is hosted on Google Cloud and functions similarly to iCloud, using Ajax.
Generative artificial intelligence chatbots can converse like people and produce new material, such as text, photos, or music, thanks to LLMs.
Ajax is said to be more capable than the prior-generation ChatGPT 3.5, and it has been trained on more than 200 billion parameters. As of September 2023, OpenAI’s newer models are allegedly more powerful than Ajax.
Apple’s Potential Generative AI Features
John Giannandrea, the chief of AI at Apple, is in charge of major language model development and answers directly to Tim Cook, the CEO of Apple. Four years ago, Giannandrea started a conversational AI team, and since then, the team’s output has increased significantly.
Although some engineers have started calling it “AppleGPT” Apple would never use that moniker for a consumer product.
The chatbot can only be used by Apple personnel with certain authorization, and its output cannot be utilized to create new features for customers’ products. It may respond to inquiries on the training data and is employed in product prototyping.
According to reports, Apple lacks a proper strategy for developing a generative AI product for consumers. Apple is experimenting with software that creates images and movies, multimodal AI technology that works with text, videos, and images, and improvements to Siri.
Below we have looked at some of Apple’s AI’s potential upcoming features that may come in handy for developers and users worldwide:
1) GPT Enabled Siri
The most evident application of generative AI, as well as the area in which it is most urgently required, is Siri, Apple’s voice assistant. With its next software updates, Apple probably will rework Siri and replace it with a chatbot that is similar to ChatGPT—a large language model (LLM) generative AI.
Bloomberg revealed in the summer of 2023 that Apple was developing an LLM chatbot internally under the moniker “AppleGPT.”
Apple wants to create a more intelligent, artificial intelligence-driven version of Siri. It intends to improve the Messages app’s interaction with Siri, making it possible for users to respond to intricate queries and better utilize sentence auto-completion.
If AppleGPT-powered Siri works as well as ChatGPT, it could enable continuous conversations—in which Siri considers previous questions to provide the most accurate response—and more comprehensive and pertinent responses.
According to rumors, Siri will be able to have more natural conversations and provide more user customization. Since the new feature will work with several devices, Siri might be able to remember conversation details across devices.
2) AI across Apple Apps
Apple’s iMovie and Photos apps are also due for an AI update, specifically in the area of visual generative AI features. It appears quite likely that users will soon be able to “zoom out” of an image by swiping away undesired elements from a photo or video or generatively filling the area immediately beyond the frame of a photo.
There are rumors that Apple wants to integrate AI into as many of its apps as possible. For instance, Apple might leverage AI to construct playlists automatically for Apple Music, and AI support in Xcode could assist developers in writing and debugging code.
AI may be used to write and create slide decks in Pages and Keynote, and it could even be modified to help with AppleCare in Apple’s internal customer care apps.
3) AI for Documentation
Extensive text messages received on an iPhone or Mac, emails, and papers could be summarised using system-wide textual generative AI. Imagine receiving an email with a 50-page PDF attachment from someone. The most important details and a brief synopsis of the PDF’s contents could be provided via the Mail app.
AppleGPT: On Cloud or Device?
AppleGPT could be available locally on user’s devices worldwide. Just a month ago, a similar approach was taken by NVIDIA’s Chat With RTX, a locally powered chatbot utilizing GPUs for high processing and querying capabilities There is a strong indication from Apple’s publicly available research paper:
“We believe as LLMs continue to grow in size and complexity, approaches like this work will be essential for harnessing their full potential in a wide range of devices and applications.”
This implies that if the new Siri chatbot is installed immediately on your Mac or iPhone, it won’t need to send your inquiry regarding a health issue to Apple’s servers for evaluation and response. Additionally, Apple never receives such information about you because it doesn’t forward your request to a server.
Similarly, your Mac or iPhone would handle the heavy job of generative AI on its own if you wanted to use it to crop a photo you shot. You wouldn’t need to submit the photo to Apple’s servers. Apple would never be able to view your edited shot or the changes you made.
Will There Be Any Possible Limitations?
Apple may need some time to update its database to accommodate new features because of the architecture of Siri.
According to The New York Times, former Siri engineer John Burkey claimed that because of the cumbersome design of the system, each time new terms are added, the database must be rebuilt, which can take up to six weeks, and more complex capabilities like search tools can take up to a year.
Apple must also address privacy issues to include generative AI features in Siri. Because Apple prioritizes privacy over functionality, it has always been more careful than its rivals. As a result, Siri has lagged behind Alexa, Google Assistant, and other voice-activated personal assistants.
When Can We Expect AppleGPT?
Analyst Jeff Pu and The Information both predict that Apple will release iOS 18 in late 2024, which coincides with the release of some sort of generative AI function for the iPhone and iPad.
Apple is constructing a few hundred AI servers in 2023, and more will follow in 2024, according to Pu’s October statement. It has been stated that Apple plans to provide both on-device processing and cloud-based artificial intelligence.
Meanwhile, you can check the comparison of GPT-4, Claude 3, and Gemini to find out which one is better.
Conclusion
AppleGPT’s release may take a lot more months, but even before its release, it has left the world of Generative AI in a frenzy. Imagine how the tool can potentially change the technical landscape with some of its features which we already discussed here. Even if rumors, this is huge news for the developer community worldwide!
