Apple is testing an AI chatbot, but doesn’t know what to do with it

Apple is developing its own AI-powered chatbot, which some engineers call the “Apple GPT,” according to Bloomberg. According to the source, the company does not yet have clear plans to release this technology to the general public.

How notes Bloomberg, the chatbot uses its own large language model (LLM) platform called “Ajax”. It runs on the Google Cloud and is built with Google JAX, a framework designed to accelerate machine learning research. Sources close to the situation say that Apple has several teams working on the project, including solving potential data privacy issues.

While other tech giants including Meta, Microsoft and Google are rapidly releasing their Generative AI products to businesses and consumers, Apple is lagging behind. Although Apple has banned its employees from using ChatGPT, Bloomberg reports that engineers are actively using the Ajax-based chatbot within the company. Ajax was created to “unify machine learning development,” says Bloomberg.

Apple has been rather modest in the field of generative AI, although the company has been incorporating AI into its software for a long time. Its most famous AI system is Siri, a voice assistant that has pioneered this field among tech companies. But critics have long argued that Siri still has a lot of room for improvement. To support its efforts, Apple hired John Giannandrea in 2018, who previously led the AI ​​and search teams at Google. He is now in charge of Siri and the machine learning teams at Apple. Giannandrea and Apple’s senior vice president of software Craig Federighi are leading Apple’s AI initiative, according to Bloomberg.

In a recent interview with Good Morning America, Tim Cook said that AI technology is something that Apple is “taking a close look at.” Cook also expressed his concerns about AI-powered products during the company’s earnings conference call in May, noting that there are “a number of issues that need to be addressed.” Other companies have begun to collaborate on the release of various LLMs for startups and researchers. Meta announced on Tuesday that its LLM LLaMA 2 will be available on the Microsoft Azure platform. Microsoft also uses OpenAI’s GPT model on its Bing search product.

While Apple’s AI plans are still unclear, sources say Apple plans to make a “significant AI-related announcement” within the next year.