The UAE Asked 11 AIs to Act Emirati. Only One Really Got It.

The UAE has just done something most countries talk about but rarely execute: it tested the world’s biggest AI models on how “Emirati” they actually are.

In a new benchmark called “AI in the Ring”, the UAE’s Artificial Intelligence, Digital Economy, and Remote Work Applications Office ranked Google Gemini 2.5 Pro as the top AI model for cultural accuracy in the UAE context. The test looked at how well AI understands local dialects, customs, traditions, and national values. 

With less than five percent of Arabic content available in digital form, and almost half of Gen Z now using AI as their main source of information, the UAE is clearly worried about who trains these systems — and on what data. 

What is the UAE’s “AI in the Ring” index?

This is the world’s first benchmark focused on how well AI models reflect Emirati identity rather than just how “smart” they are in maths or coding.

• Designed by the UAE’s AI Office as a cultural intelligence challenge
• Evaluates how AI handles Emirati dialects, traditions, values and daily life
• Built around 7 cultural dimensions and over 400 test questions
• Involved 5,200 AI-generated responses from 11 major language models 

Instead of only asking models to write code or summaries, “AI in the Ring” tests how they respond in a UAE context — from social norms and expressions to national values. A committee of Emirati experts then reviews the answers and scores each model on cultural understanding, not just language fluency. 

For readers who want to zoom out and see how Gemini is being used more broadly in the region, we’ve also covered Google’s roll-out of Gemini-powered tools in MENA in more detail on tbreak’s Google Gemini coverage. 

Google Gemini tops the cultural accuracy rankings

Gemini 2.5 Pro came out on top after all 11 models were scored on how well they align with Emirati culture and values.

• Google Gemini ranked first overall in the AI in the Ring index
• It was followed by ChatGPT (OpenAI GPT-4o) and ChatGPT (OpenAI o1)
• Cohere and Grok rounded out the top five performers
• Rankings were based on expert review of thousands of answers to culturally focused questions 

The UAE AI Office didn’t just rely on automated metrics. A committee of Emirati experts went through 5,200 responses to judge which models truly “got” the nuances of Emirati life — from how people speak, to how they respect elders, to how they talk about national identity. Gemini came out as the model that handled these layers with the highest consistency and accuracy. 

If you’re curious how Gemini is showing up in consumer products here, we’ve already seen it land inside devices like OPPO’s Reno14 series in the UAE, as covered in our Reno14 + Gemini deep dive.

Why UAE is pushing for culturally grounded AI

For the UAE, this isn’t just a tech benchmark; it’s part of a wider push to make sure AI respects the country’s identity.

• His Excellency Omar Sultan Al Olama stressed that national identity remains a priority, even as digital transformation accelerates
• He underlined that AI must reflect and reinforce authentic Emirati values
• He noted that innovation is strongest when rooted in a nation’s identity
• The goal is for AI systems to help carry UAE’s cultural heritage forward with clarity and respect 

In simple terms: if AI is going to answer questions about the UAE, it should do it in a way that feels right to people who actually live here. That includes how it talks about traditions, how it interprets social situations, and even what it chooses not to say. The AI in the Ring index is one way for the UAE to check who is doing this well — and who needs more work.

For context on how this fits into the country’s wider AI ambitions, see our feature on UAE’s big AI plans and national training push, which also involves Gemini in education.

The Arabic content gap and Gen Z’s AI dependence

The timing of this benchmark isn’t random. The UAE is clearly worried about who controls the information Gen Z sees, and in what language.

• Less than 5% of Arabic content globally exists in digital form
• Around 48% of Gen Z now relies on AI as their primary source of information
• This creates a risk that future generations learn about their own culture from models trained on non-local or inaccurate data
• UAE officials say this makes UAE-relevant cultural training data essential 

If almost half of young people are asking AI for answers, and most of the internet skews towards English content, you can see the problem. Without deliberate work like AI in the Ring, models might default to generic or foreign perspectives when responding to Arabic or Emirati-specific questions. The UAE’s move is basically a quality check before that becomes the norm.

What this means for AI users in the UAE

You won’t see “AI in the Ring” as a button in your favourite app, but its impact will show up over time in how AI behaves when you use it in the UAE.

• Government and industry now have a benchmark to compare AI models on cultural fit
• Future tenders and deployments may consider cultural alignment scores, not just price or speed
• UAE-based projects can push vendors to train on better local data
• Users may see more AI tools that “feel” local in language, tone, and context

For everyday users, this means that when you ask an AI about Emirati customs, national events, or social etiquette, the answer is more likely to respect local norms. For businesses and government, it adds another filter: not just “Is this AI powerful?” but “Does this AI actually understand us?”

FAQs

Written by

Share

Copied