Apple working on a generative AI chatbot? Heres what we know

It seems like the generative AI train is not going to stop soon even after repeated requests and warnings by technology experts and leaders around the world. Apple, which was yet to board this train, is now getting ready with its own product and services, as per a report.

Apple is working on its own artificial intelligence tools similar to ChatGPT in order to challenge OpenAI, Google and others, as reported by Bloomberg. The report says that the Apple engineers informally call the toolset ‘AppleGPT’ but its real name is Ajax because the company has used Google’s JAX framework to build its large language model (LLM).

Apple has several groups working on the project, and one group is focused on dealing with possible privacy issues.

What Apple plans to do with the bot

It is not clear what Apple will use the bot for, as the company doesn’t have any definite plans for use cases, launch dates or platforms. Bloomberg quotes a source as saying that Apple will reveal more information next year. The company also has its yearly earnings call next month where it might give some clues about Ajax.

Apple CEO Time Cook has earlier said, “We view AI as huge and we’ll continue weaving it in our products on a very thoughtful basis.” 

Bloomberg report said that Apple’s generative AI initiative is led by John Giannandrea and Apple’s senior vice president of software engineering, Craig Federighi. Giannandrea was brought in to manage Siri and its machine learning skills so we may expect some AI-powered capabilities in Apple’s digital assistant in future.

Meta’s Llama 2

Meanwhile, Meta has announced a new AI model, Llama 2. It is the successor to Llama models with improved performance over the previous generation.

Llama 2 is a set of AI models that can produce text and code based on prompts, similar to other conversational systems.

Llama 2 can be used for free for research and business purposes and can be customized on AWS, Azure and Hugging Face’s AI model hosting platform in pre-trained form. But Llama was restricted to some users as Meta decided to control access to the models because of worries about misuse.

Llama 2 comes in two forms, Llama 2 and Llama 2-Chat. Llama 2-Chat was fine-tuned for conversational exchanges. Llama 2 and Llama 2-Chat have different degrees of complexity: 7 billion parameters, 13 billion parameters and 70 billion parameters.

“Parameters” are the components of a model learned from training data and essentially determine the ability of the model on a task, in this case producing text.