OpenAI's Advanced Voice Mode is HERE! 5 New Voices, Custom Instructions & More | SEPT 2024 Update

OpenAI Finally Releases Advanced Voice Mode

OpenAI is finally rolling out advanced voice mode to all paid US and team users in ChatGPT this week. They've added several exciting improvements including custom instructions specifically for voice mode, memory functionality, and five new voices with various accents.

Real-Time Translation and New Animation

The update showcases impressive real-time translation capabilities with natural accents. In the demonstration, a user apologizes to their grandmother who only speaks Mandarin, and ChatGPT seamlessly provides the translation in the appropriate language while maintaining natural conversation flow.

There's also a new animation interface that enhances the visual experience during voice interactions.

Key Features of Advanced Voice Mode

Natural Conversations: The system now handles interruptions, humor, and more natural dialogue patterns.

Multiple Voice Options: An expanded set of voices to choose from, giving users more personalization options.

Personalized Responses: The system uses memory and custom instructions to shape responses according to user preferences.

User Control: Audio recordings are saved but can be deleted at any time, giving users full control over their data.

The Return of the "Samantha" Voice

One of the most significant updates is the return of the voice that sounds remarkably similar to Scarlett Johansson's character from the movie "Her." This voice option brings back that distinctive AI companion experience that many users were hoping for.

Custom Instructions for Voice Mode

The custom instructions feature allows users to specify exactly how they want the AI to communicate in voice mode. You can customize elements like:

• Accent and intonation preferences
• Speaking cadence (slow, clear enunciation)
• How you want to be addressed
• Preferred conversation style

In the demonstration, Charlotte, a technical program manager, shows how setting up basic information like your name and location (Bay Area) allows ChatGPT to provide personalized, location-specific recommendations for weekend activities.

Practical Use Cases

Drew, a model design engineer at OpenAI, shares how he uses advanced voice mode in his daily workflow. He keeps it running quietly beside him while working, then asks questions as they arise. These questions often develop into longer, more complex conversations.

The voice adapts to the tone of the conversation and can be prompted to assume different characters or roles:

• Interview practice partner
• Storytelling companion
• Information resource for bouncing ideas
• Character role-play scenarios

Performance and Availability

The voices demonstrate incredibly fluid conversation with low latency, making interactions feel like talking with another person rather than an AI system.

Geographic Limitations: Advanced voice mode is not yet available in the EU, UK, Switzerland, Iceland, Norway, and Liechtenstein.

Personal Perspective on the Release

While these updates are impressive, especially the custom instructions feature that allows for personality customization, it's worth noting the extended timeline. Many users, including myself, upgraded specifically for voice mode when it was first announced, only to wait months for the actual release.

This experience highlights the importance of managing expectations with AI feature announcements and considering alternative AI voice solutions that may be available in the meantime.

What's Your Experience?

Are you planning to sign up for ChatGPT's advanced voice mode, or do you already have access? How are you planning to use these new features in your workflow? The custom instructions and expanded voice options open up many possibilities for personalized AI interactions.