So, basically what the title says. I was wondering how does the speaking to ChatGPT work. Is it just a regular speech to text software like when you would say something to type a message that's been around for years now or does it have a way to analyze the sounds in the same way it was trained to analyze the text? Or is it the case that the tools for picking up speech just got better over time and it can provide more information on what it is picking up, and ChatGPT just uses that software alongside its chatbot?
I got curious about this when I heard that it can pick up accents, so it would need to provide additional information aside from just the actual words that were being said.