Apple is reportedly working on a ChatGPT-like large language model (LLM) that some within the company are calling “Apple GPT.” According to Bloomberg, Apple has built the framework to create LLM — the core of generative AI tools, including ChatGPT, Bard, and Bing.
However, there’s no clarity yet on whether Apple will release it to the public, though there could be some announcements next year. Notably, the iPhone maker restricts employees from using publicly available LMM models like ChatGPT to safeguard its sensitive data.
The report suggests that Apple has become concerned about falling behind in the AI landscape. While the company already utilises artificial intelligence and machine learning for various services, moving away from the mainstream could pose challenges, as many users are now gravitating towards generative AI tools. Apple is reportedly calling the LLM framework “Ajax,” which can provide functionalities similar to ChatGPT. This means users can ask questions and receive simplified answers on various topics.
The report highlights that the Ajax framework is built on top of Google Jax — Google’s machine learning framework. Apple already works with Google for its Google Cloud services to power Apple’s cloud network. Apple also uses its own infrastructure and Amazon Web Services (AWS). Notably, the iPhone maker also posted a job opening recently. The description notes that the candidate must have a “robust understanding of large language models and generative AI.”
Citing people familiar with the matter, the report points out that Apple started working on the Ajax framework last year. It adds, “Ajax was first created last year to unify machine learning development at Apple.” The company has also seemingly built a chatbot, though it is not widely available, even internally. For those with the bot, there’s a “special approval for access.” Overall, the Apple GPT so far remains restricted and offers no novel features.
There’s no information on how Apple is training its AI system. The company is also exploring using an enterprise version of OpenAI’s technology for employees.
While Apple barely uses terms like AI (artificial intelligence) and likes to coin its own terms to steer away from the mainstream, the company CEO Tim Cook remains optimistic about the AI space.
Last month, Cook told Google Morning America of ABC that LLM models show “great promise.” Cook is also concerned that such AI models need to be thoroughly trained to avoid bias or misinformation. The Apple chief wants self as well as government regulations. He said, “Regulation is something that’s needed in the space. I think guardrails are needed, and if you look down the road that it’s so powerful that companies have to employ their own ethical decisions. Regulation will have a difficult time staying even with the progress on this because it’s moving so quickly and so I think it’s incumbent on companies as well to regulate themselves.”
Source: India Today
