Unlocking Arabic intelligence from data to the future of AI

Empowering the Arabic Digital Future: Nour Al Hassan on Sovereign AI and Data Quality

Nour Al Hassan, CEO and founder of Arabic AI, shared her journey, starting over 16 years ago with Tarjama. This company became the largest translation provider in the MENA region. Tarjama evolved from human-centric to technology-driven in 2016, building an Arabic language ecosystem, which led to Arabic AI’s creation.

Ms. Al Hassan highlighted Arabic’s complexities: 22 dialects and 420 million speakers. Yet, Arabic content is only 4% of the internet. This digital gap causes Arabic to lag in global AI advancements, creating a critical need for specialized, localized solutions.

Launched last year, Arabic AI is a sovereign, enterprise-focused company providing full-stack AI solutions. It fine-tunes large language models (LLMs) for Arabic, progressing from a 32 billion parameter model to a more powerful 200 billion parameter model, demonstrating robust technical development.

Ms. Al Hassan emphasized Arabic AI’s focus on data sovereignty, quality, and control, not direct competition with global giants. This ensures sensitive information security for clients like banks and government entities, as generic open-source models often cannot meet these stringent privacy and integrity requirements.

Many AI deployments fail due to fragmented or unclean client data. Arabic AI addresses this by integrating AI agents directly into existing enterprise systems. This approach moves beyond isolated tools, ensuring deep integration for measurable efficiency gains and tangible business impact.

Ms. Al Hassan stressed that AI deployment must be enterprise-wide, interacting with all systems and continuously learning from proprietary data. Drawing from 16 years in translation, she affirmed the ongoing necessity of “human in the loop.” AI’s primary role is to enhance efficiency, not to immediately replace human roles.

The objective is to empower teams to achieve more, potentially doubling workload capacity, rather than increasing headcount. This requires re-skilling employees and defining desired automation levels. Overcoming employee resistance, often rooted in fear or unfamiliarity, is crucial for successful AI adoption and maximizing benefits.

Arabic AI’s comprehensive ecosystem serves diverse sectors, including banking, finance, legal, and government. It encompasses LLMs, data services (annotation, labeling, cleansing), and an “Arabic first” agentic platform. This platform features an Arabic interface, empowering Arabic speakers, including engineers, to engage more effectively with AI tools.

The company also offers an academy for skill development and a marketplace for flexible work. Arabic AI curates high-quality Arabic datasets and develops specialized models for speech and document extraction. Their collaboration with the Stanford Helm Institute for an Arabic leaderboard validates that superior data quality drives competitive AI performance.

Ms. Al Hassan concluded by advocating for innovation originating from the MENA region. With increasing government investment and a burgeoning startup ecosystem, the region is poised to adapt and localize global technological advancements. By fine-tuning AI to cultural nuances and specific market needs, Arabic AI aims to build sovereign, impactful innovation.