Apple is doubling down on enhancing the capabilities of its virtual assistant, Siri, by investing a substantial amount of resources into the development of artificial intelligence (AI) and generative AI technologies. With a renewed focus on creating conversational chatbot features for Siri, Apple is reportedly dedicating millions of dollars on a daily basis towards research and development in this domain.
Generative AI, a rapidly advancing field within AI, holds immense potential to bolster Siri’s capabilities. While Apple has not made any official public announcements about its plans, CEO Tim Cook has acknowledged the significance of generative AI and expressed his interest in the technology.
The roots of Apple’s foray into generative AI can be traced back to several years ago. John Giannandrea, Apple’s head of AI, formed a team dedicated to working on large-language models (LLMs), which serve as the foundation for generative AI-powered chatbots like ChatGPT. Currently, Apple’s conversational AI team, known as the Foundational Models team, is spearheaded by Ruoming Pang, who brings 15 years of experience from Google. Despite having a small team of just 16 members, Apple’s advancements in this field have been on par with OpenAI, a prominent AI research organization that allocated over $100 million for training a similar LLM.
According to reports, Apple has not limited its efforts to just one team. At least two other teams within the company are working on language and image models. One team focuses on Visual Intelligence, specializing in generating images, videos, and 3D scenes, while another is dedicated to multimodal AI, which can effectively handle text, images, and videos.
Apple’s immediate plan involves incorporating LLMs into Siri, enabling users to automate complex tasks using natural language, similar to the voice assistant advancements made by Google. Apple’s advanced language model, Ajax GPT, is believed to surpass OpenAI’s GPT 3.5 in terms of performance and effectiveness.
However, implementing LLMs into Apple’s products poses its own set of challenges. Unlike some competitors who rely on cloud-based approaches, Apple prioritizes on-device software processing for better privacy and performance. The main obstacle lies in the size and complexity of Apple’s LLMs, including Ajax GPT, making it challenging to fit them onto iPhones. Shrinking large models is not unprecedented, as Google has successfully developed PaLM2, which comes in different sizes, including a version suitable for devices and offline usage. Although Apple’s plans remain unclear, it is possible that the company may opt for smaller LLMs to address privacy concerns.
Q: What is generative AI?
A: Generative AI refers to the application of artificial intelligence techniques to generate or create new and original content, such as chatbot responses, images, videos, or text, without explicit programming. It involves training machine learning models to replicate human-like creativity in generating content.
Q: How does Apple plan to use LLMs for Siri?
A: Apple intends to integrate large-langua