GCC AI News — 23 Sep 2026

The UAE and Bahrain have joined an international initiative calling for stronger oversight of advanced AI, including mandatory safety testing before deployment and a possible global institution to coordinate standards. Launched on September 21 by Finland's president and Norway's prime minister during the UN General Assembly, 'A Call for Control of Frontier AI Models' has backing from 22 leaders across 20 countries plus the European Commission, with the two Gulf states the only Arab-world signatories among the initial supporters. The UAE is represented by Sheikh Abdullah bin Zayed Al Nahyan and Bahrain by Dr. Abdullatif bin Rashid Al Zayani.

It shows the Gulf helping to write the rules of responsible AI, not just consuming the technology. Being early Arab-world signatories alongside Germany, Canada and Singapore gives the UAE and Bahrain a voice in how frontier models are tested and governed worldwide, aligning ambition with public trust.

Businesses in the region can expect clearer expectations around safety testing and independent evaluation as these principles feed into national frameworks. Building documented evaluation and human-oversight steps into AI projects now positions organizations well for the standards this initiative is likely to encourage.

HUMAIN, the AI company owned by Saudi Arabia's Public Investment Fund, launched HUMAIN M3, its first frontier Arabic large language model, unveiled at the LEAP conference in Riyadh. The model uses a 428-billion-parameter mixture-of-experts architecture that activates 23 billion parameters per token, and was post-trained on more than one trillion tokens of Arabic-native content. It scored an equal-weighted average of 89.37% across seven public Arabic benchmarks, taking top marks in five. HUMAIN also unveiled HUMAIN Voice, a conversational platform supporting Saudi, Maghrebi, Egyptian and Levantine dialects, with open weights planned for release.

Arabic has long been underrepresented at the AI frontier despite hundreds of millions of speakers. A high-scoring, sovereign Arabic model developed in the Gulf means better tools for the region's own language and culture, and marks Saudi Arabia moving from AI consumer to AI producer.

Developers and enterprises across the Arab world gain a home-grown, dialect-sensitive option for chatbots, document processing and voice assistants that understand local context. Teams building Arabic-first products can evaluate HUMAIN M3 and Voice as an alternative to models tuned mainly for English.

Abu Dhabi's Executive Council has deployed an agentic AI platform for the first time to support government decision-making at its highest level. Chaired by Crown Prince Sheikh Khaled bin Mohamed on September 17, the meeting used an AI Committee System to analyse proposals, generate executive summaries and provide real-time insights during discussion. The platform ensures every output and decision remains traceable to its original source documents. It builds on Sheikh Mohammed bin Rashid's earlier pledge to deliver half of UAE government services through AI agents within two years, backed by large-scale training of government staff.

Most governments pilot AI in back-office tasks; the UAE is putting it into cabinet-level decision-making with a clear rule that every insight is traceable to a source. That combination of ambition and auditability is a model for using AI to speed up governance without sacrificing accountability.

Public-sector teams can borrow the design principle: let AI summarize and surface insights, but require every output to link back to source documents so humans can verify before deciding. It is a practical pattern for adopting agentic AI in high-stakes, accountability-sensitive settings.

Oman's Ministry of Transport, Communications and Information Technology held the inaugural meeting of a new Green Artificial Intelligence Alliance in Muscat, formed with the Ministry of Energy and Minerals and the Authority for Public Services Regulation. The alliance is tasked with promoting renewable-powered AI and high-performance computing, supporting research into net-zero computing and building national capacity in sustainable AI, in line with Oman Vision 2040. Alongside it, state-backed clean-energy platform O-Green announced plans to build more than 200 megawatts of AI-ready data-centre capacity across Oman, Spain and other markets by 2028, drawing on an 11-gigawatt renewables portfolio.

AI's soaring energy use is one of its biggest sustainability challenges. Oman pairing AI expansion with renewable power and a dedicated green-computing alliance shows the Gulf can grow its AI infrastructure while managing the environmental cost, turning abundant sun and wind into a genuine advantage.

Enterprises with sustainability commitments can look to green data-centre capacity like O-Green's to run AI workloads with a lower carbon footprint. It gives climate-conscious organizations a regional option for training and inference that aligns compute growth with net-zero goals.

The Mohamed bin Zayed University of Artificial Intelligence (MBZUAI) in Abu Dhabi has launched a long-term study, spanning up to 25 years, that uses genetic, biological and lifestyle data to advance disease prediction and preventive healthcare in the UAE. Announced in September 2026, the research aims to build AI models that identify health risks earlier and support a shift from treating illness to preventing it. The initiative draws on the region's dedicated AI graduate university and reflects a wider UAE push to apply artificial intelligence across the national healthcare system.

Preventive medicine works best with long-term, population-specific data, which most AI health tools lack. A 25-year UAE study builds exactly that foundation, and models trained on local genetics and lifestyles are more accurate for the region's people than tools built on foreign datasets.

Over time, this research could power earlier risk alerts and personalized prevention for residents, from cardiovascular to metabolic conditions. Health providers and researchers gain a regional evidence base to build screening and preventive-care tools tuned to the UAE population rather than imported assumptions.

Open Innovation AI and Dhow Engineering, part of Dhow International Holding, formed a strategic partnership in September to accelerate sovereign AI adoption in Kuwait. The collaboration combines Open Innovation AI's sovereign 'AI Fabric' technology, which orchestrates workloads across on-premises, private-cloud and hybrid environments with governance built in, with Dhow Engineering's local integration and delivery expertise. It targets government entities and enterprises in critical national sectors that need to run AI under national control, supporting use cases such as private AI assistants, document intelligence and analytics while keeping data, infrastructure and governance within the country.

Sovereign AI lets a country adopt advanced capabilities without sending sensitive data abroad. For Kuwait, building local infrastructure through partnerships, rather than depending entirely on foreign clouds, is a pragmatic path to modernizing government and enterprise services while protecting national data.

Government bodies and regulated firms in Kuwait can run private AI assistants and document intelligence on infrastructure that keeps data inside national borders. It is a template for organizations that want advanced AI but must meet strict data-residency and governance requirements.

The UAE Cyber Security Council, Open Innovation AI and Cisco, working with Emircom, announced a National AI Test and Validation Lab that lets government bodies and private organizations test, validate and certify AI models, agents and applications for security, safety and trustworthiness. Hosted in the UAE and governed by the Cybersecurity Council, the lab benchmarks against international frameworks including ISO 42001, MITRE ATLAS, NIST AI RMF and OWASP for LLMs. Assessments cover model robustness, protection against prompt-injection and jailbreaks, data-leakage and privacy risks, supply-chain integrity of models and weights, and agent-autonomy risks in tool use.

As AI spreads into public services and business, a trusted place to verify that models are secure and aligned turns AI ambition into concrete assurance for citizens. Benchmarking against global standards gives the UAE a sovereign capability to certify AI without relying on outside evaluators.

Organizations deploying AI in the UAE can have models and agents independently tested for prompt-injection resistance, data leakage and safe tool use before going live. It offers a clear certification path for building AI that regulators and customers can trust.

Qatar announced the launch of the Global Alliance for AI Ethics (GAAIE) during the inaugural session of the United Nations Global Dialogue on AI Governance in Geneva. Announced by Minister of Communications and Information Technology Mohammed bin Ali Al Mannai, the alliance, an initiative of Hamad Bin Khalifa University under Qatar Foundation, aims to strengthen international cooperation on AI ethics and responsible governance. It seeks to ensure diverse ethical values and cultural traditions shape international technical and regulatory frameworks. Al Mannai outlined priorities including bridging gaps in computing, data and talent, promoting interoperable governance, and keeping human oversight central to public trust.

Global AI rules risk reflecting only a handful of cultures. A Qatar-led alliance amplifying diverse ethical and cultural perspectives helps ensure AI governance serves the whole world, and keeps meaningful human oversight and trust at the centre of the conversation.

Policymakers and organizations can engage with the alliance's work to align AI practices with globally interoperable, culturally inclusive ethics standards. For teams operating across regions, it is a signal to build governance that respects local values rather than a single default framework.

Under its national AI policy and Economic Vision 2030, Bahrain has begun shaping an AI-embedded economy, according to an Atlantic Council analysis published in September. The kingdom has adopted an approach that emphasizes the practical application of AI across public and private initiatives, with a significant focus on expanding AI in education and partnering with other countries to equip its workforce for an AI-driven economy. Analysts also highlighted Bahrain's unique geography, surrounded by shallow waters, as a potential testing ground for innovative, climate-resilient data-centre research, positioning the country to contribute to both its AI vision and wider technology research.

Not every country needs gigawatt-scale ambitions to benefit from AI. Bahrain's emphasis on practical deployment and workforce education is a realistic model for turning AI into everyday economic value, and its focus on skills means the benefits reach people, not just infrastructure.

Smaller economies and organizations can follow Bahrain's playbook: prioritize concrete AI applications and invest heavily in training people to use them, rather than chasing scale for its own sake. Pairing adoption with education is what makes AI stick in real workplaces.

Saudi Arabia is ramping up AI and digital-infrastructure investment led by its Public Investment Fund, which anchors the strategy on investing globally, integrating AI across its own operations and accelerating adoption across portfolio companies, the Saudi Press Agency reported in mid-September. The kingdom ranked first globally in digital governance and cybersecurity indicators and third in AI models during 2025. PIF established HUMAIN to build data centres, cloud infrastructure and top-tier Arabic models while protecting local intellectual property. Internally, PIF deployed over 100 applications, activated 43 AI solutions and automated more than 1,500 processes, saving over three million working hours.

A top-three global ranking in AI models shows the Gulf is now a genuine contributor to frontier AI, not merely a buyer. Anchoring the effort in a sovereign fund that also uses AI to run itself demonstrates a full-stack commitment, from national strategy down to daily operations.

Enterprises and developers in the region gain a growing home-grown ecosystem for compute, Arabic models and enterprise AI, reducing reliance on foreign platforms. Saudi Arabia's internal automation results, millions of hours saved, offer a concrete benchmark for organizations planning their own AI-driven efficiency gains.