Pronunciator’s Use of Artificial Intelligence
Pronunciator uses artificial intelligence selectively, transparently, and responsibly. Most AI-assisted features are optional learning activities that provide useful and creative ways to interact with Pronunciator’s human-produced instructional content. AI is also used experimentally in a number of clearly identified courses for certain constructed, rare, or historical languages.
A Human-Produced Language-Learning Database
Pronunciator is the largest human-produced language-learning database in the world. Its core instructional content includes:
- More than 1 million human-translated instructional phrases:
Pronunciator’s instructional phrases have been translated by human language professionals rather than generated automatically. - More than 1 million human recordings:
Learners hear instructional phrases spoken by human speakers, providing natural models of pronunciation and language use. - Hundreds of thousands of human video recordings:
Pronunciator includes an extensive collection of video clips showing people speaking the instructional phrases. - More than 12 million human-produced audio lessons:
These lessons represent more than 6 million hours of language instruction.
Customized Audio Lessons Without Generative AI
Pronunciator’s customizable audio lessons are not generated by artificial intelligence. They are created using a proprietary audio-splicing technique developed by Pronunciator years before the recent emergence of generative AI.
This technology allows Pronunciator to assemble customized lessons from professionally prepared, human-translated, and human-recorded instructional material. Learners can personalize their lessons while continuing to hear the original human voices and instructional content.
How Pronunciator Uses AI
Pronunciator uses AI in selected activities where it can add meaningful educational or creative value. These activities provide learners with additional ways to engage with instructional material and practice the language they are studying.
For example, Pronunciator contains tens of thousands of photographs used to illustrate instructional phrases. Certain optional AI-assisted activities can analyze one of these images and:
- Create a short story:
The learner can generate a story inspired by the people, objects, setting, or actions shown in the photograph. - Generate practice sentences:
The learner can receive additional sentences based on the image and use them for reading, vocabulary, grammar, writing, or speaking practice. - Identify items in an image:
The learner can explore vocabulary associated with the objects and details visible in the photograph.
These features create new ways for learners to interact with Pronunciator’s existing instructional materials. They do not replace the human translations, recordings, videos, photographs, or lessons that form the foundation of the program.
Experimental AI-Produced Language Courses
Pronunciator also offers a small number of experimental courses for certain constructed, rare, or historical languages for which sufficient human-produced instructional material may not otherwise be available. In these particular courses, AI may be used to produce some or all of the instructional content.
These experimental languages are clearly distinguished from Pronunciator’s standard human-produced courses. In the “I want to learn” language menu, they are identified by a blue book icon rather than the normal flag icon.
The blue book icon indicates that the course is experimental and that its instructional content may have been produced primarily or exclusively with AI assistance. Because reliable linguistic resources can be limited for rare, constructed, or no-longer-spoken languages, these courses may contain inaccuracies, uncertain interpretations, or material that remains open to correction.
We welcome feedback from speakers, scholars, language communities, and other knowledgeable users. When appropriate human expertise becomes available, we are glad to review, correct, expand, or replace experimental material.
Subscribers who prefer not to make these experimental courses available to their users may have them hidden from their Pronunciator service by contacting support@pronunciator.com.
Responsible and Transparent Use
Pronunciator uses AI when it offers a useful educational capability or makes it possible to provide experimental access to languages that might otherwise be unavailable. We do not use AI merely as a shortcut for replacing the extensive human-created content at the heart of Pronunciator.
Our approach is guided by several principles:
- Human content remains central:
Pronunciator’s principal curriculum is built upon the work of human translators, speakers, educators, editors, performers, photographers, and production teams. - Optional AI activities support learning:
AI-assisted activities such as image-based stories and practice sentences provide additional forms of interaction with the underlying instructional material. - Experimental AI-produced courses are clearly identified:
Courses for certain constructed, rare, or historical languages are marked with a blue book icon rather than the standard flag icon, allowing users to distinguish them from Pronunciator’s human-produced language courses. - Institutional choice is respected:
Subscribers may ask us to hide the experimental AI-produced courses from their users. - Transparency matters:
We do not present AI-produced material as though it were created by humans, and we do not attempt to obscure how these features or courses were produced. - Educational value comes first:
AI-assisted features are evaluated according to their usefulness, creativity, accuracy, and ability to support meaningful language practice. - Corrections and human collaboration are welcomed:
We welcome input from language communities, scholars, speakers, and other qualified contributors who can help us improve or replace experimental material.
Does Bluebird Languages Use AI?
Bluebird Languages currently has no AI-assisted features or AI-produced language courses.
Like Pronunciator, Bluebird offers more than 12 million human-produced audio lessons, representing more than 6 million hours of instruction. Its customizable lessons are created using the same proprietary audio-splicing technology developed years before the emergence of modern generative AI.
Our Continuing Approach
Pronunciator will continue to evaluate new technology carefully. We believe that AI can offer valuable tools for creativity, individualized practice, and experimental access to languages for which conventional course development may be difficult.
At the same time, human language expertise remains essential to the quality and integrity of language instruction. Our objective is to combine the world’s largest human-produced language-learning database with carefully selected technologies that give learners useful new ways to practice and explore languages.