Artificial intelligence is becoming the new infrastructure of the global economy. But as countries race to build smarter systems, a deeper question is emerging: who gets to decide how machines understand language, culture, and society? For the Arab world, that question is increasingly centered on Arabic AI.

by Kasun Illankoon, Editor in Chief at Tech Revolt

[For more news, click here]

For decades, technology’s dominant languages have shaped technology’s dominant systems. English became the foundation for much of the internet, software development, and artificial intelligence research. As large language models moved from laboratories into everyday life, that imbalance became more visible: languages with massive digital footprints received stronger representation, while others often struggled with accuracy, context, and cultural understanding.

Arabic sits at the center of that challenge.

Spoken by more than 400 million people across dozens of countries, Arabic is not a single computational challenge. It is a complex linguistic ecosystem containing Modern Standard Arabic, regional dialects, historical references, cultural expressions, and legal and social contexts that vary across societies.

As governments, companies, and researchers invest billions into artificial intelligence, the question is no longer simply whether AI can speak Arabic. It is whether AI can understand Arabic-speaking societies.

For Alaa Abdulaal, Chief of Digital Economy Intelligence at the Digital Cooperation Organization (DCO), the answer depends on how countries define control over the technology shaping their future.

AI sovereignty is not about building everything alone

AI sovereignty has become one of the most discussed concepts in technology policy, but it is also one of the most misunderstood.

Some view it as a race to build every component domestically, from semiconductor manufacturing and cloud infrastructure to foundation models. Abdulaal argues that the reality is more nuanced.

“AI sovereignty is sometimes misunderstood as building everything domestically, from chips to cloud to foundation models,” she said. “In practice, in the context of Arabic AI and regional digital economies, it means having meaningful agency over the AI systems that shape national development, public services, culture, security, and economic competitiveness.”

That agency, she explained, is not defined by isolation. Instead, it is about having the ability to make informed choices.

“It means having trusted access to compute. It means being able to govern national data responsibly. It means developing or adapting models that understand Arabic, local dialects, legal systems, social norms, and cultural context.”

For Arab nations, the language dimension is particularly important.

Unlike many technology challenges that can be solved through more computing power alone, Arabic AI requires understanding the cultural and linguistic differences embedded within the language itself.

“Arabic is not one uniform computational problem,” Abdulaal said. “It includes Modern Standard Arabic, many dialects, and rich historical, cultural and social references.”

That distinction matters because a model that simply translates English-based knowledge into Arabic may produce technically correct responses while still failing to understand the people using it.

“AI sovereignty therefore includes the ability to ensure that AI systems do not merely translate from English, but understand Arabic-speaking societies and communities with greater linguistic and cultural accuracy.”

The Middle East is moving from AI consumers to AI shapers

For much of the AI boom, countries outside the traditional technology centers of North America and East Asia largely operated as consumers of imported systems.

That dynamic is beginning to change.

Across the Gulf, governments are investing heavily in AI infrastructure, cloud computing, data centers, research capabilities, and locally developed models. Saudi Arabia, the United Arab Emirates, and Qatar have emerged as some of the most visible players in the Arabic AI race.

But Abdulaal cautions against measuring progress only through the size of a model or the scale of an investment.

“The honest answer is that progress is increasing, but control remains uneven across the full AI value chain across the whole world,” she said.

Several countries have strengthened their position at the policy and adoption level, developing national AI strategies, governance frameworks, and public-sector initiatives.

However, deeper layers of the AI stack remain more complicated.

“Many AI systems used in business and government rely on externally provided foundation models, cloud platforms, chips, and technical standards,” Abdulaal explained.

That does not automatically create a problem, she added. Global partnerships are likely to remain a permanent part of AI development.

The bigger question is whether countries have the ability to understand, govern, and adapt those systems.

“The critical question is not simply ‘Is the model externally provided or locally developed?,” she said. “The better question is: can the country decide where sensitive data is processed, understand the model’s limitations, switch providers if needed, audit outcomes, enforce national laws, protect citizens’ rights, and develop local alternatives for critical use cases?”

That distinction is becoming increasingly important for governments worldwide, including in North America, where debates around AI regulation, data security, and technological dependence are accelerating.

Arabic AI has moved from an afterthought to a strategic priority

For years, Arabic representation in artificial intelligence lagged behind other major languages.

Many global AI systems were trained primarily on English-heavy datasets, creating gaps in dialect understanding, cultural awareness, and reliability in sensitive areas such as healthcare, education, and government services.

Abdulaal believes that era is beginning to shift.

“For many years, Arabic was underrepresented in global AI systems,” she said. “General-purpose models were primarily trained on English-heavy datasets, and Arabic capability was often added later.”

Now, the region is seeing the emergence of Arabic-first and Arabic-focused models designed with local needs in mind.

Saudi Arabia’s ALLaM, the UAE’s Falcon and Jais models, and Qatar’s Fanar represent a broader movement toward creating AI systems where Arabic is part of the foundation rather than an additional capability.

“These are important because they place Arabic AI from being an afterthought to being a design priority,” Abdulaal said.

Yet building models alone will not solve the challenge.

The next stage requires an entire ecosystem: better Arabic datasets, stronger benchmarks, dialect evaluation, safety testing, skilled researchers, and practical applications.

“Keeping pace with digital transformation requires more than a few strong models,” she said. “It requires high-quality Arabic data, domain-specific datasets, dialect coverage, Arabic safety benchmarks, public-sector adoption frameworks, local talent, compute access, and sustainable funding for research and startups.”

There is no single winner in the Arabic AI race

Unlike the global competition for semiconductor dominance or frontier AI models, Arabic AI is unlikely to have one clear winner.

Different countries are developing different strengths.

Abdulaal argues that ranking countries misses the larger picture because AI capability is built across multiple layers.

“Countries are progressing in different layers of the AI value chain, and from a multilateral perspective it is more useful to look at complementarity than league tables,” she said.

Saudi Arabia has focused heavily on large-scale infrastructure, investment, and national AI capabilities. Through organizations such as SDAIA, ALLaM, and HUMAIN, the Kingdom is attempting to build capabilities across data, compute, cloud, applications, and talent.

Qatar has focused on Arabic-centric AI development through projects such as Fanar, emphasizing language, culture, and multimodal capabilities.

Egypt represents another important piece of the ecosystem, with its large Arabic-speaking population, technology workforce, and ambitions around AI research, applications, and national capabilities.

Meanwhile, countries such as Bahrain, Oman, Jordan, Morocco, and Kuwait are developing important foundations through governance, skills development, startups, and sector-specific AI adoption.

The future of Arabic AI, Abdulaal argues, will come from connecting these capabilities rather than competing in isolation.

The future of Arabic AI will be shaped by an ecosystem

Five years from now, the biggest influence on Arabic AI may not come from a single company, government, or research institution.

It may come from collaboration.

“The future of Arabic AI will not be shaped by one country or one company alone,” Abdulaal said. “It will be shaped by an ecosystem.”

Governments will continue to play a central role through regulation, investment, procurement, and national strategies.

Universities and research institutions will drive breakthroughs in language understanding, safety, and evaluation.

Startups will develop applications that solve real-world problems.

Cultural organizations, universities, and communities will also play an important role in ensuring AI reflects the diversity of Arabic-speaking societies.

Global technology companies will remain part of the equation. The region will continue working with international cloud providers, chip manufacturers, and AI developers.

But Abdulaal believes the relationship must evolve.

“The decisive shift should be from passive consumption to active co-development,” she said.

Sovereignty does not mean independence from the world

One of the biggest misconceptions around AI sovereignty is that countries must choose between complete independence or complete reliance on foreign technology.

Abdulaal argues that neither approach reflects reality.

“Sovereignty is not binary,” she said. “It is not a simple yes or no. It is a spectrum of control, capability, and resilience.”

A country can still maintain sovereignty while using international infrastructure, provided it has safeguards around data, security, governance, and flexibility.

That includes trusted deployment environments, audit capabilities, cybersecurity protections, local expertise, and the ability to change providers when necessary.

But dependence without oversight creates risks.

“If a country has no visibility into the model, no control over data processing, no local talent, no local, regional, or trusted alternative pathways, and no governance mechanisms, then it is not exercising sovereignty in any meaningful sense,” Abdulaal said.

For Arabic AI, the path forward is not about rejecting global technology. It is about ensuring that Arabic-speaking societies have a meaningful role in shaping it.

The next generation of artificial intelligence will influence education, healthcare, government, business, and culture. The systems that power that future will carry assumptions about language, identity, and knowledge.

Who controls Arabic AI, then, is not simply a question about technology.

It is a question about representation, influence, and who gets to define intelligence in a multilingual world.

Related Articles:

Core42 Appoints Emma Cloney to Lead International Strategy as Europe's Sovereign AI Race Heats Up

Saudi Arabia Has Built the Right AI Foundations, According to PwC. Turning Them Into Profit Is the Next Test

Exclusive: Why Saudi Arabia's Startup Market Keeps Growing While the Rest of the Gulf Doesn't