ChatGPT's New Voice: More Human, Real-time Interaction

OpenAI is rolling out two new voice models for ChatGPT, making it sound more human and enabling real-time interaction. These models, GPT-Live-1 and a mini version, allow ChatGPT to hear and speak simultaneously. What's interesting is that this continuous interaction lets the AI receive information and generate answers at the same time. This differs from previous versions where it only responded after a user finished speaking. This capability is key for direct translation, allowing users to speak in one language and have ChatGPT translate it almost instantly. Atty Eleti, the voice product leader, says the biggest change is the more human-sounding voice, making conversations feel more natural. The system can even use filler words like "um" and "er." The bottom line is that these updates aim to make interacting with ChatGPT's voice mode much more fluid and natural, similar to how humans converse.