Apple is pouring millions daily to train its AI model
Apple is reportedly investing millions of dollars per day in artificial intelligence (AI) development, with a particular focus on conversational AI. As per The Information, the tech giant has multiple teams working on AI models, including a unit called 'Foundational Models' that is developing a chatbot for AppleCare and aims to improve Siri's ability to automate multistep tasks.
Foundational Models unit targets chatbot advancements
The 'Foundational Models' unit, led by Apple's head of AI John Giannandrea, comprises around 16 members, including several former Google engineers. Giannandrea has reportedly "expressed skepticism to colleagues about the potential usefulness of chatbots powered by AI language models." However, his team is working on creating conversational AI models that surpass the capabilities of OpenAI's GPT-3.5.
Apple believes Ajax GPT is more powerful than OpenAI's GPT-3.5
Apple's most advanced language model, known internally as Ajax GPT, has been trained on over 200 billion parameters, and it is being touted as more powerful than OpenAI's GPT-3.5. Initially created for internal use, the Ajax model remains largely restricted within the company. However, some Apple employees have access to an "Ajax" internal chatbot.
Multimodal AI and Siri's task automation potential
In addition to conversational AI, Apple is developing software to generate videos and images, as well as multimodal AI that works with images, video, and text. The company aims to create a feature allowing voice assistants like Siri to automate multi-step tasks, which could potentially be integrated into future Apple products.
