We present BAREC-10M, an expanded version of the Balanced Arabic Readability Evaluation Corpus (BAREC). This new release extends the original 1M-word corpus to 10 million words and broadens its scope to include balanced multi-domain coverage annotated for morphology, syntax, and readability. The corpus integrates 45 sub-corpora drawn from diverse sources, including news, educational materials, literature, children’s texts, and religious discourse. Each text is labeled for domain, readership level, and genre, and automatically analyzed using state-of-the-art morphological and syntactic tools. To enhance coverage of underrepresented varieties, we manually digitized and included children’s materials, magazines, and curriculum-based content. The resulting dataset provides a balanced resource for studying Arabic linguistic variation across styles, audiences, and levels of complexity.

Complex linguistic phenomena such as stereotypes or irony are still challenging to detect, particularly due to the lower availability of annotated data. In this paper, we explore Back-Translation (BT) as a data augmentation method to enhance such datasets by artificially introducing semantics-preserving variations. We investigate French and Italian as source languages on two multilingual datasets annotated for the presence of stereotypes or irony and evaluate French/Italian, English, andArabic as pivot languages for the BT process. We also investigate cross-translation, i.e., augmenting one language subset of a multilingual dataset with translated instances from the other languages. We conduct an intrinsic evaluation of the quality of back-translated instances, identifying linguistic or translation model-specific errors that may occur with BT. We also perform an extrinsic evaluation of different data augmentation configurations to train a multilingual Transformer-based classifier forstereotype or irony detection on mono-lingual data.