Mozilla Common Voice is an open-source initiative led by the Mozilla Foundation to build publicly available voice datasets that help improve speech recognition technology.

Unlike proprietary datasets, Common Voice is designed to be:

• Open and accessible

• Community-driven

• Representative of diverse languages, accents, and speaking styles

These datasets are used by researchers, developers, and organizations around the world to train and improve speech AI systems.

By contributing sentences and recordings, participants help create more inclusive and accurate voice technologies that better represent real people.

Learn more: https://commonvoice.mozilla.org/

Effect AI collaborates with the Mozilla ecosystem by contributing structured datasets created through the Effect Alpha workforce.

These contributions help:

• Expand available training data

• Improve linguistic diversity

• Support open AI development

Explore Effect AI’s dataset contributions:

https://datacollective.mozillafoundation.org/datasets/cmkfm9fbl00nto0070sdcrak2

Your participation directly supports real datasets that become part of the broader open voice ecosystem.

The Mozilla Data Collective is a public platform where datasets are shared and documented for transparency and collaboration.

It provides:

• Access to open datasets

• Visibility into contributors and projects

• Resources for researchers and developers building speech technologies

Explore the Data Collective:

https://datacollective.mozillafoundation.org/

This test evaluates your ability to create high-quality sentences suitable for speech datasets.

Focus areas include:

• Text generation quality

• Linguistic integrity

• Policy compliance

• Readability and natural speech

These standards help ensure contributed sentences are useful, safe, and representative of real-world language.

The assessment consists of 30 total questions covering:

• Sentence Purpose & Dataset Integrity

• Originality & Copyright

• Sentence Structure & Length

• Readability & Speakability

• Numbers, Symbols & Formatting

• Grammar, Clarity & Neutrality

• Bias, Safety & Content Restrictions

• Language Consistency & Locale

• Dataset Diversity & Usefulness

• Final Quality Judgment

Estimated duration: 10–15 minutes.

Please review carefully before starting.

Timed Questions

• Each question includes a countdown timer.

• If time runs out, your current answer is automatically submitted.

• The assessment advances to the next question.

Forward-Only Navigation

• Once a question is submitted, you cannot return to it.

• Previous answers cannot be changed.

To pass the assessment:

• You must score at least 80% (24 out of 30 correct answers).

You have 3 total attempts for this assessment.

Important:

• Starting the test counts as one attempt.

• Interrupted sessions still count toward attempt limits.

• Additional attempts are not available once all attempts are used. The Effect AI Team cannot reset your attempts.

Before starting:

• Use a stable internet connection.

• Avoid switching devices or browsers during the test.

• Ensure you can complete the assessment without interruption.

Step 1: Start the Assessment

Begin from the Capability Marketplace.

Step 2: Answer Questions

• Read each scenario carefully.

• Apply Common Voice contribution guidelines.

• Work steadily within the timed format.

Step 3: Evaluation

After completion:

• Your responses are evaluated based on dataset guidelines and policy standards.

Step 4: Capability Unlock

If you pass:

• The Common Voice Contributor capability may be granted.

• Sentence contribution workflows may become available.

Step 5: Access New Tasks

After unlocking:

• Return to your dashboard.

• Check for newly available Common Voice contribution tasks.

Once the assessment is completed:

• Your score is calculated.

• Eligible capabilities can be claimed.

• Access updates may appear shortly afterward.

If something seems unclear:

• Review the Common Voice task guides.

• Check Discord announcements.

• Ask questions in Alpha support channels.