magic starSummarize by Aili

How the voices for ChatGPT were chosen

๐ŸŒˆ Abstract

The article discusses the process of selecting the five distinct voices used in ChatGPT's Voice Mode feature, which was launched in September 2023. It covers the collaboration with the voice acting industry, the criteria used to select the voices, the audition process, and the plans for expanding Voice Mode in the upcoming GPT-4o release.

๐Ÿ™‹ Q&A

[01] Voice Mode Selection Process

1. What was the process for selecting the five voices used in ChatGPT's Voice Mode?

  • The process involved working with industry-leading casting and directing professionals to narrow down over 400 submissions before selecting the 5 voices.
  • The casting agency and OpenAI's casting directors issued a call for talent in May 2023, receiving over 400 submissions from voice and screen actors.
  • An initial list of 14 actors was independently reviewed and hand-selected by the casting team, which was further refined before presenting the top voices to OpenAI.
  • OpenAI's internal team reviewed the voices from a product and research perspective, and the final voices for Breeze, Cove, Ember, Juniper, and Sky were selected.

2. What criteria were used to select the voices?

  • The criteria included:
    • Actors from diverse backgrounds or who could speak multiple languages
    • A voice that feels timeless
    • An approachable voice that inspires trust
    • A warm, engaging, confidence-inspiring, charismatic voice with rich tone
    • Natural and easy to listen to

3. How were the selected actors involved in the process?

  • The selected actors flew to San Francisco for recording sessions and in-person meetings with the OpenAI product and research teams.
  • OpenAI discussed the vision for human-AI voice interactions, the technology's capabilities and limitations, and the safeguards implemented with each actor.
  • The actors have continued to collaborate with OpenAI, contributing additional work for audio research and new voice capabilities in GPT-4o.

[02] Upcoming Voice Mode Enhancements

1. What new features are planned for Voice Mode in GPT-4o?

  • A new Voice Mode for GPT-4o will be made available to ChatGPT Plus users in the coming weeks.
  • The new Voice Mode in GPT-4o will offer a more natural interaction, handling interruptions smoothly, managing group conversations effectively, filtering out background noise, and adapting to tone.

2. What are the plans for introducing additional voices in ChatGPT?

  • OpenAI plans to introduce additional voices in ChatGPT to better match the diverse interests and preferences of users.
Shared by Daniel Chen ยท
ยฉ 2024 NewMotor Inc.