Apple is spending millions of dollars per day to bring generative AI capabilities to Siri, says report

livemint.com:

Apple is reportedly spending millions of dollars on research and development daily in order to boost the conversational features of Siri. Reportedly, Apple is working on multiple AI models across various teams. According to a report by The Information, Apple's unit that is working on Conversational AI is called Foundation models while the company has created two other teams that are working on developing image or language models.

Apple's visual intelligence unit is working on generating images, videos and 3D scenes, similar to the likes of Midjourney or OpenAI’s Dall-E 2. Meanwhile, another team is working on multimodal AI which could recognize and produce images, video and text. Apple's artificial intelligence efforts are being headed by John Giannandrea who was hired by the company in 2018 in a bid to improve Siri.

The Cupertino-based company's advanced LLM, AjaxGPT is reported to be trained on over 200 billion parameters and could be even more powerful than GPT-3.5. The report notes that one of these models could be used for creating a chatbot aimed at users who use AppleCare. Meanwhile, the company is working on developing features for Siri that will allow iPhone users to automate tasks that involve multiple steps using a simple voice command.

Apple's voice assistant could be used to create a GIF by utilizing the last five photos taken and sending it to a friend. According to a report by Bloomberg, Apple began laying the foundation for AI services with the Ajax framework last year, as well as a ChatGPT-like tool for use internally. Ajax was created to unify machine learning development at Apple, according to the people familiar with the effort.