Apple’s commitment to advancing Siri’s conversational prowess is undeniable. The tech giant is making substantial investments in research and development, with a strong focus on various AI models. These efforts span multiple teams within Apple.
The division dedicated to conversational AI is aptly named “Foundation Models,” and it’s only one part of a more extensive initiative. Apple has also established two additional teams specializing in image and language models.
The visual intelligence unit is primarily engaged in the creation of images, videos, and 3D scenes—a task reminiscent of impressive projects like Midjourney and OpenAI’s Dall-E 2. Concurrently, another team is working diligently on multimodal AI, capable of understanding and generating images, videos, and text.
An exciting development is the potential creation of a chatbot, catering specifically to AppleCare users. This chatbot could offer enhanced customer support and assistance.
But that’s not all. Apple is tirelessly working on Siri features that empower iPhone users to streamline complex tasks with simple voice commands. For instance, Siri might soon assist in crafting GIFs using the most recent five photos taken, and then send them to friends seamlessly.
These ambitious AI endeavors are under the leadership of John Giannandrea, who joined Apple in 2018 with the mission of enhancing Siri. One standout achievement is Apple’s advanced LLM (Language Model), called AjaxGPT. This model has been trained on an astounding 200 billion parameters, potentially making it even more robust than OpenAI’s GPT-3.5.
According to insights from Bloomberg, Apple initiated its AI journey by laying the groundwork last year. The Ajax framework was introduced, along with a tool similar to ChatGPT, designed initially for internal purposes. Ajax serves as a unifying platform for machine learning development within Apple, streamlining their AI efforts.
In summary, Apple’s ongoing investments and dedicated teams are shaping a future where Siri becomes not just a voice assistant but a versatile AI companion, capable of assisting users across various tasks and interactions.
