Abu Dhabi Arabic Language Centre and NYUAD Launch AI Readability Tools in BAREC 2.0

The ALC is part of the Department of Culture and Tourism – Abu Dhabi. BAREC 2.0 moves the project beyond building a corpus. It adds AI tools that rewrite Arabic texts to match a chosen reading proficiency level and explain why a text was classified at a given level.

The agreement behind the project was signed earlier. It was announced at the fifth Congress of Arabic and Creative Industries at Manarat Al Saadiyat, with ALC chairman Dr Ali bin Tamim and NYUAD vice chancellor Professor Fabio Piano in attendance.

From Corpus to Rewriting Tools

The ALC started BAREC in 2023 with researchers from NYU Abu Dhabi and Zayed University. The aim was a 10-million-word Arabic corpus covering many genres, subjects and countries of origin, annotated by readability level. The first phase produced more than 69,000 sentences, according to the media office.

The new phase builds on that data. Its AI models adjust readability by changing vocabulary and sentence structure, either to simplify a passage or to rewrite it at a more advanced level. Because the system explains each classification, it is easier to see how it reaches its judgements, and the project's backers say this also makes it more useful in the classroom.

The partners said the tools are meant for writers producing material for different audiences, for teachers preparing resources for learners at different levels, and for students who want to simplify hard texts and build their comprehension.

Bin Tamim said BAREC 2.0 combines linguistic expertise with advanced AI models to give educators, writers and students practical ways to assess and adapt Arabic texts. Piano said NYU Abu Dhabi was "extremely proud to lead BAREC 2.0". He added that the results will help teachers choose and adapt materials and strengthen curriculum design.

Why It Matters

Arabic remains under-served in AI tools compared with English, especially in education, where graded reading materials take a lot of work to produce by hand. A large annotated readability corpus, combined with models that rewrite text and explain their output, gives schools, publishers and edtech developers resources that are scarce in the language.

The project is also another example of Abu Dhabi using its universities and AI institutions on Arabic-language technology, alongside government-backed efforts to build Arabic large language models and datasets. MBZUAI's involvement links the readability work to the emirate's main AI research university.

The announcement did not say when the BAREC 2.0 tools will be available to the public, how they will be distributed, or how much the project costs.