Skip to main content

Research & classroom

The classroom exchange behind Live Linguist

The language faculty at Kennesaw State University and Université Hassan II Casablanca conceived an international virtual exchange and wrote its research proposal. The proposal included an AI-supported caption simplifier. Dylan Goldblatt, who worked in the KSU Department of World Languages and Cultures and founded Saga Labs there, built Live Linguist with his own resources after leaving KSU for a corporate AI/ML engineering role.

The founding deployment

The first international team of linguists to deploy Live Linguist was the KSU and Hassan II team named below. The language faculty conceived the exchange and wrote its research proposal, “Language Education and Intercultural Connections with AI-Supported International Virtual Exchanges,” and submitted it to the French in Higher Education Grant Program of the Albertine Foundation. The proposal did not win support. The team designed the exchange, the curriculum, and the classroom deployment. Dylan built the software with his own resources as an independent scholar; it remains a hobby side project alongside his corporate engineering work.

The pedagogical brief

Comprehensible input, kept in the target language

Live Linguist produces a verbatim transcript and a simplified transcript. The simplified version brings spoken French at CEFR C2 down to a level a beginner can follow, CEFR A1 to B2, while staying in French. The learner gets comprehensible input in the target language.

Stephen Krashen’s Input Hypothesis holds that we acquire language through input we can understand, pitched one step beyond our current level and often written as i+1. Input beyond that level becomes noise, and input below it teaches nothing new. Simplification lowers the complexity while keeping the learner inside the language they came for.

Easy-language standards serve accessibility too. Leichte Sprache, FALC, Lectura Fácil, Selkokieli, Lättläst, Yasashii Nihongo, and the pan-European Easy-to-Read framework help people with cognitive disabilities and low-literacy readers take part in public life. A simplifier that follows those standards serves readers in their own language.

The rationale rests on second-language-acquisition theory and established accessibility standards. It makes no clinical claim about learning outcomes. Automatic register compliance supports human easy-language review; it does not replace it.

The people who conceived the exchange

The exchange team

The language faculty below designed the exchange, the curriculum, and the classroom study. Dylan built the software on his own.

Kennesaw State University

Department of World Languages and Cultures

  • Dr. Abigail Alexander. Associate Professor of French; Director of the World Languages Resource Collection. Project coordinator; conceived and has led the KSU–Hassan II virtual exchange since fall 2023.
  • Dr. Noëlle Lively. Senior Lecturer of French and Coordinator of French. Curriculum design and exchange coordination.
  • Dr. Federica Santini. Chair, Department of World Languages and Cultures.
  • Brooke Reed. Program Manager, World Languages Resource Collection. Coordinated student testers.

Université Hassan II Casablanca

Casablanca, Morocco

  • Dr. Meriem Hachimi. Associate Professor of French. Co-coordinates the exchange and Moroccan student participation.
  • Dr. Abdelhadi Samadi. Associate Dean for Research and Cooperation. Partner-institution authorization.

Independent development

Dylan Goldblatt, Ph.D. worked in the Department of World Languages and Cultures at Kennesaw State University, where he founded Saga Labs. He proposed Live Linguist as part of the exchange’s research effort. That proposal did not win support, so he built the project with his own resources as an independent scholar after starting a corporate career as an AI/ML engineer. He designed the fine-tuned easy-language Qwen3 models, the supervised dataset, the deterministic register validators, the on-device applications, and the CEFR-level caption adaptation. Live Linguist remains a hobby side project alongside his corporate work on translation, transcription, and knowledge work. He maintains the models and infrastructure.

An invitation to the field

Send requests and links to your work

The models and dataset are common tools. The next useful version will come from people who know the learners, the registers, and the public services involved. We ask four groups to write in.

Language educators

Tell us the language, the register, the learner level, and the classroom moment where a simplified transcript would help. Send a lesson or activity you would want to run.

Interpreters & accessibility teams

Tell us what “easy” has to mean in your language and setting, and where a machine draft would save time without replacing human review.

Public-sector officials

Tell us which documents, services, or public meetings need easy-language versions, and which standards your office is accountable to.

Builders & researchers

Share what you build with the open models or dataset, such as an app, a lesson, an evaluation, or a new register, so the next team can start from your work.

Details to send

  • The language and easy-language register, if you know it.
  • The learner or reader group, level, and setting.
  • The platform you use, and whether on-device processing is required.
  • A link to anything you have built, taught, or published with the tools.

We cannot promise every request ships. We read every message, and requests that fit the model line and register work have the best chance of becoming a release.

References

Scholarly context

  • Krashen, S. D. (1985). The Input Hypothesis: Issues and Implications. Longman.
  • Inclusion Europe. (2009). Information for All: European standards for making information easy to read and understand.
  • International Organization for Standardization. (2023). ISO 24495-1:2023. Plain language, Part 1: Governing principles and guidelines.
  • Xu, W., Napoles, C., Pavlick, E., Chen, Q., & Callison-Burch, C. (2016). Optimizing statistical machine translation for text simplification. Transactions of the Association for Computational Linguistics, 4, 401–415.
  • Popović, M. (2015). chrF: character n-gram F-score for automatic MT evaluation. Proceedings of the Tenth Workshop on Statistical Machine Translation.

Use, citation & ownership

Free for education, with citation samples and clear ownership

Dylan Goldblatt created Live Linguist’s free software, open-source datasets, and open-source models for educational purposes. They support language learners, educators, accessibility teams, and researchers. They do not replace a certified accessibility service, a medical device, or human easy-language review.

If you use the project in teaching, research, or a published build, attribute Dylan Goldblatt. The samples below cover the usual style guides; update the version and year when you cite a later release. For the models and dataset, cite the Hugging Face repository you used.

APA 7

Goldblatt, D. (2026). Live Linguist: On-device easy-language caption simplification (Version 3) [Computer software]. https://ndgold.com/live-linguist/

MLA 9

Goldblatt, Dylan. Live Linguist: On-Device Easy-Language Caption Simplification. Version 3, ndgold.com, 2026, ndgold.com/live-linguist/.

Chicago 17

Goldblatt, Dylan. Live Linguist: On-Device Easy-Language Caption Simplification. Version 3. Computer software. 2026. https://ndgold.com/live-linguist/.

IEEE

D. Goldblatt, "Live Linguist: On-device easy-language caption simplification," Version 3, 2026. [Online]. Available: https://ndgold.com/live-linguist/

BibTeX: project

@software{goldblatt2026livelinguist,
  author  = {Goldblatt, Dylan},
  title   = {Live Linguist: On-Device Easy-Language Caption Simplification},
  year    = {2026},
  version = {3},
  url     = {https://ndgold.com/live-linguist/},
  note    = {Open models and datasets for 29 easy-language locales}
}

BibTeX: quality model

@misc{goldblatt2026easylanguage,
  author    = {Goldblatt, Dylan},
  title     = {Qwen3-1.7B-EasyLanguage-4bit},
  year      = {2026},
  publisher = {Hugging Face},
  url       = {https://huggingface.co/ndgold/Qwen3-1.7B-EasyLanguage-4bit}
}

Ownership

Aquilops LLC holds the sole intellectual property in the closed-source applications. They cost nothing, and their source stays closed. © 2026 Aquilops LLC. All rights reserved.

The open models and datasets carry their own licenses: Apache-2.0 for the models, CC BY-SA 4.0 for the dataset. The Qwen3 base model and German4All source data keep their upstream licenses. Aquilops LLC owns the closed-source application, its name, and its icon.