Description
Summary:
Babel Audio seeks expressive English-speaking voice actors to record scripted, two-person conversations for training AI models, focusing on delivering strong emotions.
Highlights:
1. Perform as a User or AI Assistant in scripted emotional conversations
2. Bring big, clearly audible emotion to life from a script
3. Record conversations at your convenience
About Babel Audio
Babel Audio partners with leading technology companies to collect high-quality voice data that trains frontier AI models. We work with thousands of participants worldwide to capture authentic human conversation, so that AI can better understand and communicate with humanity.
About the Role
Babel Audio is looking for expressive English-speaking voice actors in the US and Canada for a scripted, two-person project. You and a partner will record a conversation together over a call. One of you plays a person going through a real moment (the "User"). The other plays an AI voice assistant (the "Assistant") who helps them with a real-world task and "uses tools" out loud along the way ("let me pull up your order," "I'll set a reminder for that").
What matters most is how things are said. The User moves through strong emotions as the scene unfolds, such as sad, frustrated, relieved or euphoric, and the Assistant responds the way a good friend would. We're looking for very emotive performers who can bring a script to life and are comfortable going big on emotion. Big, heavy emotion is the goal. Too much is better than too little: we can coach you to dial it back, but we can't add feeling that isn't there.
You might play:
A caller panicking because their flight was cancelled on the way to a wedding
Someone thrilled, then nervous, as they budget their first paycheck
A pet owner trying to stay calm while booking an urgent vet visit
An assistant who stays warm and steady with an upset caller, and laughs along at a joke
Scripts are provided, so no writing or improv is needed. Each conversation runs about 5 to 10 minutes. You choose when to record.
Who Can Apply?
Native or fully fluent English speakers living in the United States or Canada
People who can deliver strong, clearly audible emotion from a script, from joy and relief to frustration and fear
People who can bring big, clearly audible emotion while sounding like they're really talking with a partner (not reading), react in the moment and keep the energy up for a full conversation
Voice-acting or acting experience is a plus, not a requirement
People who own a quality external microphone or headset (see Technical Requirements below)
Compensation
$75 USD / recorded hour
Additional performance-based incentives may be available
Technical Requirements
Clear and consistent voice quality
External microphone: USB preferred (or XLR; but no built-in, wireless/HF, or Bluetooth microphones)
Recording setup: Computer required (no mobile devices or smartphones)
Environment: Quiet, no background noise, acoustically treated, or soundproof space (no noticeable echo)
Internet connection: Stable connection for high-quality recording
Pop filter or foam windscreen: Strongly recommended
Apply now