Voice Over Artist: AI Voice Data Training

Overview

Voices is looking to hire fluent speakers for two one-off freelance projects to record voice samples in the following languages:

  • English
  • French (France)
  • Spanish (South America)
  • Portuguese (Brazil)
  • German
  • Italian
  • Arabic
  • Hindi
  • Bengali
  • Vietnamese
  • Thai
  • Indonesian
  • Tagalog
  • Japanese
  • Russian
  • Korean
  • Mandarin


Selected candidates will be asked to record at least 1 Hour of audio on their own time, from home in their most fluent (first) language. The recordings will be used strictly for internal text-to-speech (TTS) research and development.

This is an AI Training opportunity where contributors help develop conversational AI models by recording scripted speech through our platform. To participate, applicants must create an account on our platform, where all recordings and communication will take place. This is an ongoing project rather than a one-off recording session. We welcome applicants who are fluent in the required language(s). As part of the application process, you may be asked to provide a short audio sample (usually under one minute) to confirm recording quality.

Please note:

  • These are one-time freelance engagements with no follow-up work required. No prior voice acting experience necessary, though it is welcome.
  • These postings are for project based client engagements facilitated by Voices. These are not internal Voices positions.

Key Responsibilities:

  • Record provided scripts in a quiet environment.
  • Deliver high-quality audio files in accordance with audio specifications.

Requirements & Qualifications:

  • Native or fluent in one of the listed languages
  • Access to a quiet space for high-quality voice recording
  • Access to high-quality recording equipment (a microphone and soundproofing are necessary).
  • Ability to meet the following audio specifications:
    • Minimal background noise and echo
    • Format: WAV (Waveform Audio File Format)
    • Sample Rate: 44.1 kHz or higher (48kHz is great)
    • Bit Depth: 16-bit or higher (24-bit is great)
    • Noise floor: Less than -60 dBFS (Studio-quality)
    • No background noise
    • Submission reflects raw, unprocessed audio — no reverb, compression gate, noise reduction, or any other effects.
  • Available to record approximately 1 finished hour of audio using a provided script (9000 words) and deliver audio files within 2 days of your hiring date.
  • Previous experience in voice acting, narration, or broadcasting is required.
  • Talent must be a native speaker of the posted language.
  • You will be required to sign various agreements, including an NDA and a usage release agreement.

Artistic Direction:

DO’s:

  • Be Yourself! Speak in your most natural, conversational voice.
  • Embrace Emotion!
  • Don’t be afraid to exaggerate emotions.

DON’Ts:

  • Don’t ad-lib, please read the script exactly as it is written
  • No Robot Voices!, No flat, neutral, or robotic delivery. Do not do an impression of a speech assistant like Siri or Alexa
  • Avoid Commercial Vibes! This isn’t an ad or an explainer, so no overly polished or “announcer” tones.
  • No Impressions! Just use your own unique voice.
  • Don’t record the Tone, Emotion, Character, Direction, etc. columns in the script. Please only record the ‘Script Content’ column.

Compensation:

Pay: $120-400 USD per hour, depending on language & location.

Payment Details:

  • Payment will be issued upon completion.
  • You must have a valid PayPal account or the ability to receive payment via Tipalti.
  • If using Tipalti, a verified bank account is required to process your payment.

Licensing & Usage:

  • No public or commercial use
  • Internal research and development use only
  • Recordings will not be sold, broadcast, or used to train public-facing AI

Recordings collected for this project will be used for evaluation purposes only. Your voice samples will be used to prompt speech-enabled AI systems and to assess the accuracy, safety, and quality of the AI’s responses.

These recordings will not be used to train new AI models, develop voice synthesis, or create digital voice replicas. Your voice will not be cloned, generated, or used in synthetic speech.
The purpose of this project is to test and measure existing AI systems, ensuring that they respond appropriately, safely, and responsibly in a wide variety of situations. All data will be stored securely and handled in compliance with privacy and data protection standards.


Audition Process / Apply below:

Interested? Submit your audition at the following links:

Project 1:

Project 2:

We’ll review all submissions and contact selected candidates with the next steps. AI-generated content is not permitted; all submissions must consist solely of original, human performances. Submitting AI auditions or multiple auditions in an attempt to be hired more than once is strictly prohibited. Files will be reviewed and authenticated manually and by detection software. If a talent is found to have used AI or was hired more than once due to multiple audition submissions, they will not be compensated for any of their work, their files will be deleted, and they will be removed from this project and future projects.

Schedule:

  • Flexible
  • Remote
  • One-time recording session, asynchronous

Work Location: Remote / From home

Voices is committed to collecting AI data ethically, with full speaker consent and fair compensation.

Continued opportunities:

Even if you’re not selected this time, submitting your first project sets you up for recurring work in your language as new matching Voice Data opportunities become available, and you join Voices.com’s roster of remote Voice Data contributors.

To learn more, visit Our Commitment to Ethical AI.
https://www.voices.com/blog/our-commitment-to-ethical-ai/

Thank you,

– The Voices.com Team