Amidst the wave of AI advancements and applications flooding the market, Apple continues to push forward. The company, which has been integrating AI features into its products for years, is developing its own version of a large-scale language model, similar to OpenAI’s ChatGPT and Google’s Bard. This project is internally known as “Apple GPT.”
The information leaked by Mark Gurman in Bloomberg reveals that Apple has built its own artificial intelligence framework called “Ajax” for creating these large language models. The project has gained momentum in recent months and has already yielded its first results with the development of a chatbot service that some engineers are already referring to as “Apple GPT.”
Apple Ajax can change everything
The sudden interest in artificial intelligence can be attributed to the impact that solutions like ChatGPT and, to a lesser extent, Google Bard have had in the market. These generative tools have the ability to create essays, images, and even videos based on simple text prompts. The potential of these systems has captivated both users and companies and has led to a surge in related products.
Apple has been integrating AI features into its products for years, such as computational photography, computational audio for noise reduction, intelligent adaptive mode, plant and animal recognition systems in the Photos app, and more recently, the new autocorrector in iOS 17. These are noteworthy and highly useful projects, but they may not have the same astonishing effect as launching a description and having an artificial intelligence deliver the finished photo.
Tim Cook has always been cautious when commenting on the wave of new AI services. His stance has oscillated between recognizing the potential of this technology and acknowledging the areas that still need improvement. He expressed this during an internal Apple meeting focused on AI.
Meanwhile, it’s clear that these generative tools have a lot of potential, and Apple is evidently observing the landscape while developing its own solutions. According to Gurman, Ajax is the backbone of Apple’s AI project and was initially created to unify machine learning development within the company.
It seems to have already brought improvements in areas like Siri and search, but it’s now being used to create a language model that serves as the basis for an internal tool, much like ChatGPT. This tool, as Gurman highlights, is expanding within the company, but its responses are not to be used to develop customer-facing features.
For now, it appears that employees are using it for prototyping tasks and to summarize text and answer questions based on the data with which it has been trained.
A core project that can lead to GPT
As per the information available to us, Apple’s chatbot essentially replicates the capabilities of Bard, ChatGPT, and Bing AI, without distinguishing itself from any of them in terms of functionality. The tool, without a defined design, is available as a web application for exclusive internal use. As far as we know, Apple has no plans to release it to the public, although it is continuously improving the models that power it.
Apple’s AI project, Ajax, spans across the entire company and involves software engineering teams as well as the cloud engineering team, which provides the necessary infrastructure. Various initiatives are being considered, stemming from these advancements. According to Gurman, these initiatives could lead to a significant launch in 2024.
Apple’s progress in this field is driven by vice presidents like John Giannandrea, the head of machine learning and AI, and Craig Federighi, Apple’s senior vice president of software engineering. These executives fuel the debate on how to proceed in this field, with Giannandrea advocating a cautious approach, emphasizing the need to monitor market developments.
Like other Apple cloud services, Ajax relies on Google Cloud infrastructure, particularly Google Jax, as well as Amazon AWS and Apple’s own servers. This approach allows Apple to access the necessary power and scalability while retaining control of sensitive information on its own servers.
Indeed, we may be just a few months away from Apple unveiling Apple GPT or enabling Siri to summarize entire web pages or virtually answer any question we have. The development of these different projects remains to be seen, but one thing is evident: we all had assumed that Apple was actively working on AI, and now it’s becoming a reality. We can look forward to being surprised by what the company has in store for us in this field.
Some of the links added in the article are part of affiliate campaigns and may represent benefits for Softonic.
