Apple develops AI rival to counteract Google and OpenAI

Apple is actively engaged in developing artificial intelligence tools that have the potential to compete with offerings from competitor Alphabet's Google Bard and creator of ChatGPT, OpenAI.

The focus of their efforts lies in creating large language models similar to ChatGPT and Google's Bard, using an in-house framework named "Ajax." Notably, Apple has also crafted a chatbot service internally referred to as "Apple GPT," as reported by Bloomberg News.

Apple's internal tool—developed using Ajax— is said to be similar to existing solutions like Bard, ChatGPT, and Bing AI. It currently functions as a web application with a stripped-down design, not intended for public use. As of now, Apple has no plans to release this tool to consumers, although they are actively working to enhance the underlying models.

The report says that Apple is still exploring potential consumer angles for its applications. The company is pursuing various related initiatives, with collaboration between its AI and software engineering groups, along with the cloud services engineering team, which would provide the necessary infrastructure for significant new features.

Although Apple has yet to finalize a concrete plan, the report states that the company aims to make a notable AI-related announcement in the coming year.

Meanwhile, even Meta, formerly known as Facebook, is making a significant move towards open-source AI in order to catch up with its peers after Google and Microsoft made significant strides into the AI space. It has introduced LLaMA 2, a large language model that is available for public use at no cost.

LLaMA 2 is the follow-up to Meta's earlier open-source AI model, LLaMA, which gained recognition as a valuable tool for academic research. This new iteration benefits from a 40% increase in training data and over a million human annotations to enhance the quality of its output. In a strategic move, Microsoft will be distributing LLaMA 2 through its Azure cloud service, specifically optimized for the Windows operating system.

(This story has been updated to reflect changes in the headline and correct some factual errors.)