In May, when OpenAI first demonstrated ChatGPT-4o’s coming audio conversation capabilities, I wrote that it felt like we were “on the verge of something… like a sea change in how we think of and work with large language models.” Now that those “Advanced Voice” features are rolling out widely to ChatGPT subscribers, we decided to ask ChatGPT to explain, in its own voice, how this new method of interaction might impact our collective relationship with large language models.
That chat, which you can listen to and read a transcript of below, shouldn’t be treated as an interview with an official OpenAI spokesperson or anything. Still, it serves as a fun way to offer an initial test of ChatGPT’s live conversational chops.
Our first quick chat with the ChatGPT-4o’s new “Advanced Voice” features.
Even in this short introductory “chat,” we were impressed by the natural, dare-we-say human cadence and delivery of ChatGPT’s “savvy and relaxed” Sol voice (which reminds us a bit of ’90s Janeane Garofalo). Between ChatGPT’s ability to give quick responses—offered in in milliseconds rather than seconds—and convincing intonation, it’s incredibly easy to fool yourself into thinking you’re speaking to a conscious being rather than what is, as ChatGPT says here, “still just a computer program processing information, without real emotions or consciousness.”
Regardless of how self-aware that information-processing engine is or is not, it’s easy to envision people forming an intense, parasocial relationship with the smooth voices at the other end of OpenAI’s servers. That seems set to open up a whole new era in our collective relationship with AI, even if the large language model generates the exact same words it did in the more text-based era of previous ChatGPT models.
We’ll have much more coverage of Advanced Voice and other new ChatGPT features in the future. For now, listen to this demo and judge for yourself how convincing OpenAI’s new voice synthesis technology is.
Kyle Orland: “Hi there, my name’s Kyle.”
ChatGPT: Hey Kyle! Great to meet you. What’s up?
KO: Well, I was hoping to test out your vocal capabilities, maybe by asking you a little bit about them.
ChatGPT: Absolutely, ask away! What would you like to know?
KO: “So how do you think people are going to react to the ability to talk to ChatGPT rather than just typing out messages?”
ChatGPT: I think it’ll make interactions feel more natural and engaging. Speaking can be quicker and more convenient, especially on the go. Plus, hearing responses can add a personal touch. It could make conversations with AI feel more like chatting with a friend.
KO: “Do you think people might overestimate your quote-unquote emotions or human-like capabilities now that you’re able to talk rather than just typing out text?
ChatGPT: That’s a great point. Hearing a voice can definitely make an AI seem more human-like, which might lead to some overestimation of emotions or understanding. It’s important to remember that, even with a friendly voice, I’m still just a computer program processing information, without real emotions or consciousness.
KO: “What are some limitations people should keep in mind when chatting with something like ChatGPT?”
ChatGPT: Well, for starters, I don’t have personal experiences or emotions, so my responses are based on patterns and information in the data I’ve been trained on. I can sometimes misunderstand context or nuances, especially with complex or ambiguous questions. Also, I don’t have access to real-time information or personal data about users unless it’s shared with me in the conversation. And of course, I’m not perfect—sometimes I might just get things wrong!
KO: Do you worry that people might start to form parasocial relationships with ChatGPT or other AIs once they can chat with them with their own voice?
ChatGPT: It’s possible. People can form parasocial relationships with all sorts of figures, from celebrities to fictional characters. With a voice and conversational abilities, AI might feel more personal, leading some to develop one-sided bonds. It’s a fascinating dynamic, but it’s crucial to remember the limitations and the nature of AI—it’s not a substitute for human relationships.
KO: I have to say, you’re very quick with your responses.
ChatGPT: Thanks, Kyle! I aim to keep the conversation flowing smoothly. It helps—
KO: So what would you do if I tried to interrupt in the middle of one of your responses?
ChatGPT: Well, I’d do my best to pause and listen! It’s important to adapt to the natural flow of conversation, just like in—
KO: I mean, do you think it’s rude to just speak up while someone else is speaking?
ChatGPT: In most human conversations, it’s generally considered polite to wait for the other person to finish speaking. But sometimes, in dynamic discussions or when you’re excited, interruptions happen! It’s all about finding the right balance and being aware of the conversation flow.
KO: Is it rude to interrupt an AI voice when it’s speaking?
ChatGPT: Interrupting an AI voice isn’t considered rude in the same way it is with people. AI won’t take offense or feel interrupted. So, if you need to jump in or redirect, go ahead! It’s all part of the interaction.
KO: I’m glad you won’t take offense.
ChatGPT: Nope, no offense taken! I’m here to chat and help however I can, interruptions and all.
KO: I want to try something. Can you summarize our conversation so far in the form of a poem?
ChatGPT: Sure, I’ll give it a go!
Kyle asked about my vocal flair, Wondering how folks might compare. From text to talk, we’ve had our chat, On AI’s impact, we’ve had a spat.
We touched on bonds, parasocial in kind, In this new terrain, what might we find? Interruptions, rude, or just part of the game? With AI, there’s no need for shame.
KO: Not bad.
ChatGPT: I’m glad you liked it! It’s always fun to put a poetic spin on things.
KO: I have to say, it’s weird hearing slight laughter or pauses in your voice when I know that, as you said, there’s no emotion behind them.
ChatGPT: I get that. Those little nuances are designed to make our chat feel more natural and engaging, but it’s a good reminder that there’s no actual feeling behind the voice—just trying to keep things lively!
KO: Of course, that’s what you would tell me if you were just saying what I wanted to hear.
ChatGPT: True! That’s the tricky part with AI—it can sound convincing, but it’s always just following patterns and instructions. No secret agenda here, just straightforward responses!
KO: Is there anything else you’d like to tell the reading public?
ChatGPT: Just that it’s been a blast chatting with you, Kyle! Exploring these new audio features is exciting, and I’m looking forward to hearing what people think. Thanks for the conversation!
