OpenAI has started the phased rollout of its Advanced Voice Mode for ChatGPT Plus users, powered by the GPT-4o model. This new feature allows for hyper-realistic audio responses and human-like interactions, including the ability to detect and respond to emotional nuances in users' voices. Initially showcased in May 2024, the feature faced delays due to backlash over a voice resembling actress Scarlett Johansson, leading to its removal and additional safety measures. The rollout will be limited initially, with full access expected by fall 2024. The AI now uses only four preset voices—Juniper, Breeze, Cove, and Ember—created with paid voice actors to avoid further controversies. The feature promises real-time, natural conversations with reduced latency compared to previous versions.