Apple Hopes To Launch Own ‘Apple GPT’ AI Model, Codenamed ‘Ajax’: Report

Tech giant Apple has been conspicuously absent from the rush to launch large language model programs like OpenAI’s ChatGPT, but that could soon change according to a Bloomberg report. 

On Wednesday, the site reported that Apple has a framework in place to develop its own model, with the codename “Ajax,” which will be the basis for a program that Apple engineers are currently playfully calling “Apple GPT.”

The report cited anonymous sources within the company who said that several teams are working on the project, though no timeline for a public release was given. 

Apple did not immediately respond to The Messenger's request for comment.

Large language model chatbots have become a hot item thanks to the success of ChatGPT, with Google launching Bard in February and Microsoft unveiling its Bing AI the following month. Apple has been unusually quiet on this front, even as other companies like Meta and Elon Musk’s xAI have rushed to make announcements about their own intentions. On Tuesday, Meta went so far as to make its LlaMA 2 AI open source, meaning anyone can build a chatbot off the technology. 

Apple hasn’t been entirely absent, though it has focused on features that are able to run on the relatively limited processing power of a single device, versus the thousands of GPUs required to run a program like ChatGPT. In June, the company announced several programs that would use artificial intelligence and machine learning, including an updated autocorrect for iPhones.

“In those moments where you just want to type a ducking word, well, the keyboard will learn it, too,” said Craig Federighi, Apple’s chief of software, at the time. 

During an earnings call in May, Apple CEO Tim Cook didn’t mention AI during his prepared remarks but during a question period said the company needed to be “deliberate and thoughtful in how you approach these things.”