Official announcements or press releases from major tech companies

OpenAI Unveils GPT-4o 'Omni' Voice Model and Voice Engine at NYTechWeek

Jun 8, 2024, 02:48 PM

OpenAI introduced their new GPT-4o voice model at the NYTechWeek event. The GPT-4o, also known as 'omni,' is designed to talk, see, and understand like a human, with potential applications in various industries such as edtech and robotics. The model utilizes a text-to-speech (TTS) system that generates human-like audio from text using a 15-second sample of speech and employs a diffusion process to match the speaker's voice. OpenAI has also published an update on the Voice Engine, detailing its functionality and safety research, addressing concerns related to the Sky voice. OpenAI has engaged with global policymakers to discuss the potential and risks associated with the technology.