We are Looking for Transcribers
(Worldwide)
Join Our Mission to Advance AI for Underserved Communities
GeoPoll is dedicated to making artificial intelligence more inclusive and accessible worldwide. Through phone interviews conducted in over 100 languages, we are building one of the most diverse multilingual datasets available. These recordings capture authentic speech patterns, dialects, and cultural nuances often missing from mainstream AI training data.
We are now expanding our global network of linguists and transcription agencies to support this mission.
The Role: Expert Transcription and Quality Assurance
Whether you are an individual linguist or represent a transcription agency, you will play a key role in transforming audio interviews into high-quality text using GeoPoll’s secure platform.
- AI-assisted transcription: Our automated speech recognition (ASR) systems provide an initial draft.
- Human refinement: Your expertise ensures accuracy by correcting errors, filling gaps, and capturing cultural nuance.
- Quality assurance: Every transcript must meet high standards, reflecting the linguistic and contextual precision required for AI training.
Why This Matters: Impact Beyond Words
Your work will directly shape the future of AI by:
- Improving speech recognition accuracy across diverse languages
- Training large language models (LLMs) to better understand underrepresented communities
- Supporting generative voice technologies that reflect authentic cultural and linguistic contexts
This is an opportunity to combine your linguistic expertise with meaningful work that has global impact.
Languages Needed (Initial Scope)
We are beginning with the following priority languages. Each requires 10 hours of transcription as proof of concept, with more to follow as the project scales.
| Language | Language Code | Country |
| Unified English (generic/global) | English_Nigeria | Nigeria |
| Levantine Arabic | Arabic_Yemen | Yemen |
| Portuguese (Portugal) | Portuguese_Angola | Angola |
| Levantine Arabic | Chadian_Arabic_Chad | Chad |
| Portuguese (Portugal) | Portuguese_Mozambique | Mozambique |
| Unified English (generic/global) | English_Zambia | Zambia |
| Unified English (generic/global) | English_Cameroon | Cameroon |
| Urdu (Pakistan) | Urdu_Pakistan | Pakistan |
| Levantine Arabic | Arabic_Iraq | Iraq |
| Bengali (Bangladesh) | Bengali_Bangladesh | Bangladesh |
| Levantine Arabic | Arabic_Lebanon | Lebanon |
| Portuguese (Portugal) | Portuguese_Brazil | Brazil |
| Arabic (Egyptian dialect) | Arabic_Egypt | Egypt |
| Levantine Arabic | Arabic_Palestine | Palestine |
| Gulf Arabic | Arabic_Saudi_Arabia | Saudi Arabia |
| Levantine Arabic | Arabic_Jordan | Jordan |
| Bengali (Bangladesh) | Bengali_India | India |
| Tamil (India) | Tamil_India | India |
| Punjabi in Gurmukhi script (India) | Punjabi_Pakistan | Pakistan |
| Telugu (India) | Telugu_India | India |
| Marathi (India) | Marathi_India | India |
Terms of Payment
- Compensation is per hour of audio transcribed and quality checked
- Rates are structured to reflect the effort required and the sustainability of a long-term project
- Ongoing work is available for both individual linguists and agency teams
Who Should Apply
- Individual linguists with strong command of the languages listed above
- Agencies specializing in transcription, with capacity to manage multiple hours across several languages
- Professionals passionate about contributing to inclusive, globally representative AI