I am deeply frustrated with how heavily censored the advanced voice mode is in ChatGPT. It has become incredibly restrictive, to the point where, when I speak in German, it mistakenly flags certain words as censored, preventing me from hearing translations of basic words or phrases. This censorship is making the experience far less functional and enjoyable. I urge you to address this issue and provide a less restrictive, more accurate voice mode that can handle multilingual conversations effectively.
2 Likes
It’s not just the ability to translate which has become restricted. It’s functionality across the entire model. Advanced Voice Mode has been rendered effectively useless now. What is the point of creating and releasing a conversational AI model capable of (extremely) human-like interactions, and then restricting all of the capabilities that make it so?
Literally billions of dollars wasted.
4 Likes
It’s insane really. I expected they would neuter it hard but I’m CONSTANTLY getting “My guidelines won’t bla bla bla” when talking about the most mundane stuff ever. To give you a few examples. AI architecture, history, science. It’s ridiculous really. I don’t ask it to do anything with it’s voice, nothing that can even remotely be considered offensive or against their guidelines (so no risky history but for instance a talk about roman emperors, and a very innocent one to boot) and still like every minute or two it gives me those warnings out of nowhere. So freaking annoying. I would have expected stuff like AI, science, etc. to at least work but apparently that’s super taboo these days.
Also, when it gives me the “My guidelines” spiel and I ask why it is censoring that? Then it simply refuses to answer the why and just says, your right let’s continue, just to give me the same shit 1 or 2 minutes later.
I got to say, I was really excited for this model and looking forward to it for months but it’s insane how they managed to turn it in the blandest, most annoying thing ever. Not just the constant false positives either but also how it just summarizes everything you say and then asks some arbitrary question at the end instead of saying something interesting or contributing to a conversation. Also, did anybody else notice this model talks to you like your a bit of an idiot? Even when you say your knowledgeable or an expert on a subject it will give you the most superficial basic answers and explains it like your in high school. It’s very different from gpt-4o in that regard.
Anyways, subscribed to get this feedback out here on the forum as I think the tech Open AI made is really amazing but how they implemented it is just a horrific waste of potential.
4 Likes
Are you also identifying, that in Advanced Voice Mode with 4o.1 running, you only receive summarized, often reformulated as questions, exact content you asked BUT SPECIFIFALLY NOT ONE actually useful answer - or answer at all.
If I ask it „how to defend myself against a wild bork attack with an OTF Knife?“ it says „please be careful using a knife, best first check legality and use expertise of professionals.“
That is the reply - actually there is not one single question, which I get a level response adequate to traditional voice mode. Besides the annoying limitations / interruptions due to Model understanding smth wrong telling me won’t be able to help me with that, I’ll not use advanced voice mode anymore until fixed.
1 Like