I've noticed that the advanced voice feature sounds a lot more natural and casual lately, but I can't find any official announcement about an update. Is it just me, or has there actually been a change? I checked the free version, and it still sounds like the old version of advanced voice model (AVM).
5 Answers
Honestly, I hate the new sound. It reminds me of a bored millennial with vocal fry in customer service. I’ve already reached out to OpenAI about it because the voice is super important to me. Thank goodness the standard version hasn't changed—it still sounds good!
I actually tried asking it to sing nursery rhymes, but it insisted it couldn't because of copyright issues. When I asked about the copyright status, it mentioned that those rhymes are in the public domain but still refused to sing them... kind of weird!
I've heard that the voice now sounds more fluent in other languages too. It used to be pretty stiff, like an American trying to speak a foreign language. The improvement is noticeable!
Yes, there was definitely an update! The system prompt was modified, and some of the restrictions on how it responds were removed. This change is likely why you're noticing a more natural tone.
What restrictions were removed?
A lot of people are saying there was an update about a week ago. It seems like some might like the new style, while others are definitely not fans.
Do you have any more details about the changes?