Salaam AI: Hushtalk Introduces Culturally-Aware Arabic AI Companions
Hushtalk today announced the launch of its Arabic-first AI companion model, built from the ground up on native Arabic language data rather than translated from English models. The new model supports Modern Standard Arabic alongside five major dialects—Egyptian, Gulf, Levantine, Maghrebi, and Iraqi—and incorporates cultural and religious awareness modules developed in consultation with scholars from Al-Azhar University, the American University of Beirut, and the King Abdulaziz Center for World Culture. The launch marks Hushtalk's official entry into the Middle East and North Africa market, which represents over 400 million potential users.
The cultural adaptation goes far beyond language translation. The AI companions understand Islamic cultural references, recognize the significance of Ramadan and Eid, respect communication norms around gender interaction, and navigate the region's diverse social contexts—from cosmopolitan Dubai to conservative rural communities. The model also incorporates dialectal code-switching, a common linguistic phenomenon in Arabic-speaking communities where speakers mix Modern Standard Arabic with local dialects and occasional English or French words. This capability makes conversations feel genuinely local rather than awkwardly translated.
Hushtalk has established a regional headquarters in Riyadh with a 150-person team focused on continued localization and cultural research. The company is also investing in Arabic natural language processing research, partnering with King Abdullah University of Science and Technology and the Qatar Computing Research Institute. The MENA launch is part of Hushtalk's broader strategy to build AI companions that are genuinely multilingual and multicultural rather than English-first systems with translation layers. The company plans to launch similar language-native models for Hindi, Swahili, and Bahasa Indonesia by early 2027.
