OpenAI has recently shared the behind-the-scenes process of selecting voices for ChatGPT, shedding light on the careful planning and collaboration with top casting professionals.

The introduction of voice capabilities in ChatGPT back in September 2023 was a game-changer, offering users a new way to interact with AI. The five distinct voices—Breeze, Cove, Ember, Juniper, and Sky—were carefully chosen after an extensive five-month search that involved reviewing over 400 submissions.

It’s worth noting that Sky’s voice is not an imitation of Scarlett Johansson but belongs to a talented actress using her natural speaking voice. To bring ChatGPT’s ‘Voice Mode’ feature to life, OpenAI worked closely with the voice acting industry, ensuring that the selected voices met high standards of quality and appeal. The company collaborated with well-known casting directors and producers to set criteria for the selection process.

In early 2023, OpenAI set out on the quest to find the perfect voices. They focused on finding actors who could offer diverse backgrounds, multilingual capabilities, and a timeless, engaging voice. By May 2023, over 400 submissions poured in from voice and screen actors within just one week.

The casting team meticulously reviewed these submissions and narrowed down the pool to 14 promising candidates. These voices were then presented to OpenAI for further evaluation. Detailed discussions ensued between OpenAI and each actor about the vision for ChatGPT’s Voice Mode, ensuring that they understood the project’s scope and intentions before committing.

By June, the final voices were selected, and OpenAI brought the actors to San Francisco for collaboration. This partnership went beyond voice recording as actors also contributed to audio research and new voice capabilities for GPT-4o.

Finally, on September 25th, 2023, these voices were officially launched in ChatGPT for users’ enjoyment.