Apple has significantly increased its investment in artificial intelligence (AI) over the past few months, focusing on conversational AI, automation, and multimodal systems, according to a new report from The Information.
In July, it was reported that Apple was working on its own framework as well as a Large Language Model (LLM) that has the potential to power what Apple employees are calling “Apple GPT”.
The new report shows that Apple now spends millions of dollars per day training LLMs under its Foundational Models team, which was established four years ago and is led by Siri’s chief, John Giannandrea.
Despite Giannandrea’s skepticism towards chatbots, Apple has been experimenting with its own internal bots like Ajax, which is said to be more advanced than OpenAI’s ChatGPT 3.5.
However, with the launch of chatbots like Microsoft’s Bing and Google’s Bard, Apple has yet to unveil any consumer-facing conversational AI products. The 16-person Foundational Models team pales in comparison to the resources dedicated by Apple’s competitors.
Nonetheless, Apple appears to have wide-ranging AI ambitions beyond chatbots. One project aims to develop a Siri feature that can automate multi-step tasks voiced by users, eliminating the need for iOS Shortcuts workflows. This could arrive as early as iOS 18, according to sources.
Other teams are working on AI-generated video and imagery and multimodal systems that can understand relationships between text, images, and video. Apple is also researching ways to improve privacy protections for its AI models.
Apple is determined not to fall behind as the AI race heats up in Silicon Valley. However, a beloved analyst, Ming-Chi Kuo, said last month that Apple’s progress in the AI industry is “significantly behind its competitors.”
For example, Microsoft has invested heavily in this area. The company has integrated AI into many of its products. That includes Bing Chat AI and its Copilot features coming to Microsoft 365 and Windows 11, among other products.
Source: The Information