Saudi Arabia is advancing globally in artificial intelligence standards

Saudi Arabia has achieved advanced global ranks in axes and standards related to artificial intelligence, among 70 countries, according to the results of the global digital competitiveness and artificial intelligence classification, issued by the International Institute for Management Development (IMD), on Wednesday.

Saudi Arabia came second globally in the two criteria: companies’ awareness of the cybersecurity risks associated with artificial intelligence, and protecting sensitive data in companies’ systems.

The results of the classification showed that Saudi Arabia ranked fourth globally in three criteria, which included implementing the national strategy for artificial intelligence, adopting it by companies, and making its technologies available to members of society.

The Kingdom ranked fifth globally in the axis of adopting artificial intelligence, and in three sub-criteria, which included the level of citizens’ confidence in it, the regulatory frameworks for issues related to it, and companies’ investments in it.

Saudi Arabia ranked seventh globally in three criteria: artificial intelligence skills, cities’ use of it to improve the quality of services, and the extent to which companies address bias in its algorithms, while it ranked eighth globally in the axis of readiness for it.

Private sector investments in artificial intelligence

The Kingdom ranked ninth in the world in the criterion of private sector investments in artificial intelligence, and eleventh in the world in the axis of investment in artificial intelligence and capital.

These results reflect the progress that Saudi Arabia is making in developing the data and artificial intelligence system, enhancing its readiness and expanding the scope of adopting its technologies in various sectors, in addition to developing national capabilities, stimulating investments and developing regulatory frameworks, in a way that enhances its global competitiveness and contributes to achieving the goals of “Saudi Vision 2030.”

On the other hand, a study issued by the Saudi Data and Artificial Intelligence Authority (SDAIA) reviewed the Kingdom’s efforts and progress in the field across 7 basic pillars and 15 sub-pillars, including policies and regulations, investment, infrastructure, data, human talent and capabilities, research, development and innovation, and institutional readiness and adoption, highlighting the country’s progress in global indicators and the development of its national capabilities.

A study entitled “The State of Artificial Intelligence in the Kingdom” highlighted Saudi Arabia’s progress in global indicators, by ranking third in the world in the rate of growth of artificial intelligence competencies according to the Stanford Institute index for the year 2026, and third in the world and first in the Arab and regional countries in developing leading models according to the same index for the year 2025, and fifth in the world and first in the Arab world in sector growth according to the “The Observer” index for the year 2025, and obtaining first place in the world after achieving 26 Award in the 2025 Global Youth Artificial Intelligence Competition, and in 2026, it will join the Global Partnership for Artificial Intelligence (GPAI).

Infrastructure and data

The study revealed an increase in the total capacity of data centers in the Kingdom by 115.8 percent between the years 2023 and 2025, and the high-performance computer “Shaheen 3” topped the Gulf countries according to the list of the best 500 high-performance computers in the world for the year 2026, along with the “Hexagon” data center, the largest government data center in the world with a Tier IV classification and a total capacity of 480 megawatts, in addition to the presence of 9 cloud regions in the Kingdom until mid-2020. 2026, including 4 regions under construction developed by global cloud service providers.

The study showed the integration of data from more than 515 government systems into the National Data Lake of the National Data Bank, with a hosted data volume exceeding 5.87 petabytes, provided by about 90 government agencies during the year 2026, while the number of open data sets exceeded 21.8 thousand sets published by 346 government and private agencies during the same year, in addition to the use of more than 500 billion Arabic data units to develop the “Allam” linguistic model (ALLaM 34B) in the year. 2025, and the total linguistic and phonetic words on the “Falak” platform of the King Salman International Academy for the Arabic Language reached 2.2 billion words during the same year.

The pillar of talent, research and innovation

The study showed that 84 percent of the public in Saudi Arabia have knowledge of the concept of artificial intelligence, and 74 percent have awareness of its real-life applications, while the total beneficiaries of the “SMAI” initiative in its first edition exceeded 1.2 million as of July 2026, and the number of graduates holding technical degrees in specializations including computer science and information technology reached more than 69.3 thousand graduates between 2021 and 2025, with a compound annual growth rate recorded of 15 per cent. 100 for the density of artificial intelligence talent in the Saudi market during the same period.

The study monitored an average annual growth rate of 37.4 percent in the number of scientific publications specialized in artificial intelligence between 2021 and 2025, and 34.52 percent in the number of patents between 2020 and 2024, in addition to the development of 8 large language models and 58 local artificial intelligence applications, according to the annual report of the Ministry of Communications and Information Technology for the year 2026, and the development of 5 pioneering models within the Kingdom according to the Stanford Institute index for the year 2025.

Institutional readiness and adoption

كشفت الدراسة عن أن 90 في المائة من الجهات الحكومية في السعودية تمتلك استراتيجية للذكاء الاصطناعي وإطار عمل موحد لقياس نجاح مشاريعها وأثرها، فيما تستخدم 97 في المائة من الجهات الحكومية نماذج الذكاء الاصطناعي وأدواته أو تطورها، وتعتمد 77 في المائة منها ممارسات لحوكمة الذكاء الاصطناعي، بينما تلتزم 63 في المائة بوضع إجراءات رسمية للامتثال للمبادئ الوطنية الصادرة عن «سدايا»، في حين سجلت 48 في المائة من الجهات الحكومية التي تتبنى التقنيات زيادة في كفاءة عملياتها الداخلية تتجاوز 10 في المائة.

The study reviewed the Kingdom’s advanced ranks in specialized global indicators within four axes: إذ حققت في محور البيئة التنظيمية المرتبة السادسة عالمياً في الاستراتيجية الحكومية، والسابعة في الحوكمة، والثالثة في مرصد سياسات الذكاء الاصطناعي، والثانية عشرة عالمياً والأولى محلياً وإقليمياً في سلامة الذكاء الاصطناعي، فيما جاءت في محور رأس المال البشري بالمرتبة الأولى عالمياً في تمكين المرأة في المجال، والثالثة في نسبة الطلاب المستخدمين للذكاء الاصطناعي التوليدي، والرابعة في استقطاب الكفاءات.

وحلت السعودية الأولى عالمياً في الأمن والخصوصية والتشفير في الذكاء الاصطناعي، والأولى في الجاهزية الرقمية، والثامنة في الاستشهادات العلمية بالذكاء الاصطناعي، فيما حققت في محور التنافسية الاقتصادية المرتبة السابعة عالمياً في تمويل القطاع المالي، إلى جانب المراتب المتقدمة في نمو الذكاء الاصطناعي وتطوير نماذجه الرائدة ونمو كفاءاته، التي استعرضتها الدراسة ضمن مؤشرات تقدم المملكة عالمياً.

وتأتي هذه الدراسة ضمن جهود «سدايا» في إبراز تقدم السعودية في مجال البيانات والذكاء الاصطناعي، وتسليط الضوء على المنجزات الوطنية والمراكز المتقدمة في المؤشرات العالمية، والتعريف بالقدرات والإمكانات الوطنية في مجالات البنية التحتية والبيانات والمواهب والبحث والابتكار والتبني المؤسسي، مما يعكس تطور منظومة الذكاء الاصطناعي في المملكة ونمو قدراتها.

تتصدر مطابقة المعاملات بين الشركات التابعة والتسويات المصرفية وعمليات الإقفال المالي، قائمة المجالات التي تحقق فيها الإدارات المالية السعودية فوائد ملموسة من الذكاء الاصطناعي والأتمتة، وفقاً لأنور حديدي، مسؤول النمو في قطاع التدقيق والضمان، ومسؤول قطاع التدقيق في السعودية لدى «ديلويت» الشرق الأوسط.

ويوضح حديدي، في مقابلة مع «الشرق الأوسط» أن هذه العمليات تتميز بكثافة المعاملات ووضوح إجراءاتها، ما يجعلها أكثر ملاءمة للأتمتة، بينما لا تزال تطبيقات أكثر تعقيداً، مثل التنبؤ المالي، تواجه تحديات تتعلق باتساق النتائج وقابليتها للتفسير وثقة الإدارات بها قبل التوسع في استخدامها.

وتعكس هذه الفجوة نتائج تقرير «اتجاهات المالية 2027» الصادر عن «ديلويت»، الذي أظهر أن 42 في المائة من قادة المالية المشاركين في استطلاع تناول طموحاتهم في الذكاء الاصطناعي الوكيل وجاهزيتهم للتوسع حتى عام 2027، يرون أن طموحاتهم تتجاوز القدرات المتاحة حالياً، بسبب الحاجة إلى ضوابط أقوى قبل توسيع نطاق الاستخدام.

ويشرح حديدي أن القيمة الأكثر وضوحاً في السوق السعودية تظهر في عمليات مطابقة الحسابات والإقفال المالي، لأنها تعتمد على معاملات كبيرة الحجم وبيانات منظمة نسبياً. أما التنبؤات المالية، فعلى الرغم من إمكاناتها، لا تزال مؤسسات كثيرة تعمل على إثبات اتساق نتائجها وقابليتها للتفسير، وبناء ثقة الإدارة بها، قبل عدّ فوائدها قابلة للتوسع والاعتماد عليها بصورة منتظمة.

The gap between experience and actual operation

يوضح حديدي أن أحد الأسباب الرئيسية لتعثر الانتقال من التجارب الأولية إلى الأنظمة التشغيلية هو تصميم كثير من مشاريع الذكاء الاصطناعي لإثبات الإمكانات التقنية، بدلاً من اختبار قدرتها على العمل بصورة موثوقة ضمن العمليات المالية اليومية.

فقد تحقق التجربة نتائج جيدة اعتماداً على بيانات اختيرت ونُظمت مسبقاً، لكنها تواجه تحديات مختلفة عند التعامل مع بيانات المؤسسة الفعلية، بما تتضمنه من استثناءات وتفاوت في الجودة وتعقيدات تشغيلية.

Some experimental applications also operate in isolation from the core financial management systems, without a clear definition of who is responsible for the results, required controls, or escalation procedures when errors or unexpected results are discovered.

For the CFO, the model's ability to perform the task in specific test conditions is not enough. فالثقة التشغيلية تتطلب دمج النظام في سير العمل، واختباره بصورة مستمرة، وتحديد المسؤوليات المتعلقة بالمخرجات والاستثناءات والقرارات التي تتأثر بها.

Hadidi points out that, before expanding use, organizations need to determine who owns each process, the controls that govern the automated workflow, and how exceptional cases are transmitted to employees authorized to review them and make decisions regarding them. These requirements become important when technology moves from providing ancillary analyzes to producing information that is used directly in financial reports or management decisions.

Who is responsible for AI decisions?

As financial departments increasingly rely on analyzes produced by models, the issue of responsibility for decisions based on these analyzes arises. Hadidi believes that responsibility should remain with management, and not with technology, even when artificial intelligence plays an important role in preparing a recommendation or analyzing data.

This means that the competent financial official is still required to review the results, understand the basis on which they were based, and bear responsibility for the final decision and its consequences. Therefore, the cases in which the use of artificial intelligence is permitted, the level of human review required, the entity authorized to approve the results, and the mechanism for documenting approvals should be determined so that they can be verified later.

Hadidi stresses that speeding up processes should not come at the expense of oversight, stressing the need for outputs to be traceable to their original data, the logic used in producing them, the record of model versions and modifications, and management reviews. It also calls for these applications to be subject to change management controls, access permissions, exception handling, and periodic testing, similar to other financial systems. The standard to be applied to financial results is summarized by saying that if management cannot interpret a certain number, it should not be included in official reports.

Automating repetitive work and the limits of professional judgment

According to Hadidi, the processes most suitable for automation currently include accounts payable, invoice matching, reconciliations, expense verification, and processing routine accounting entries. These areas are characterized by repetition of tasks and the presence of clear rules for their implementation, which allows for reducing manual intervention while maintaining oversight and review mechanisms. However, this does not mean that the same approach can be applied to all financial functions, as the need for professional judgment remains essential in operations that involve complex accounting estimates or interpretations.

The most prominent examples include assessing impairment of assets, establishing provisions, estimating expected credit losses, recognizing revenue in complex cases, and evaluating an entity's ability to continue its business. In these areas, AI may provide information and analysis that supports professionals, but it does not replace their responsibility to evaluate assumptions, understand the business context, and make accounting decisions.

In turn, technology can enhance financial oversight by examining entire data sets, rather than limited scopes, and early detection of unusual accounting entries or violations of separation of powers and approval procedures. However, these capabilities come with risks that include changes to models, unauthorized access to information, weak oversight, and over-reliance on automated results. This is why Hadidi stresses the need to integrate artificial intelligence into the existing internal control system, and not allow it to work outside it.

A data problem that models don't solve

Hadidi describes data quality as the most important factor in the success of financial AI, and at the same time one of the aspects that organizations often underestimate the challenges associated with it. Many organizations use different enterprise resource planning systems, non-standardized structures of accounts, benchmark data of varying quality and different definitions of indicators used in reports. These differences lead to difficulties in collecting and analyzing information consistently, especially when results depend on data from multiple subsidiaries or departments.

Hadidi cautions against believing that investing in more advanced models will address basic data problems, explaining that artificial intelligence may uncover these problems and amplify their effects rather than solve them. He notes that a frequent mistake is purchasing advanced tools before establishing clear rules for data governance, determining its ownership, sources, and transmission paths, and working to unify them. These foundations become necessary when an organization moves from using limited tools to applications directly related to preparing accounts, forecasts, and reports.

Governance privacy in the Saudi market

Data protection and cybersecurity requirements add considerations to decisions to adopt artificial intelligence in Saudi Arabia that go beyond the efficiency of the tools and their technical capabilities. Hadidi explains that organizations need to evaluate the transparency of models, mechanisms for dealing with sensitive information, and governance of cloud applications, especially when processing operations involve transferring or dealing with data across borders. These complexities arise in large groups operating across multiple companies and in different jurisdictions, where determining where data is located and who is authorized to process and access it becomes part of the design of the financial system itself.

Hadidi links these challenges to the importance of data residence requirements, the need to consider sovereign artificial intelligence solutions, in addition to strengthening the role of boards of directors and audit committees in supervising the risks and controls associated with the technology. The matter is not limited to choosing the technical environment in which the models are run; Rather, it extends to ensuring that the organization is able to document processing operations, manage access permissions, track results, and verify the safety of their use.

This underscores the importance of coordination between financial management, technology, security, risk and internal audit teams, rather than dealing with artificial intelligence as a separate project managed by one entity.

How is the actual return measured?

Hadidi does not believe that the time saved by automation alone is a sufficient measure of the success of investing in artificial intelligence. A more accurate assessment should include shortening and improving the reliability of financial close cycles, reducing manual intervention, reducing adjustments resulting from audits, and improving the accuracy of forecasts. This adds to the speed of processing exceptional cases, reducing the cost of a single financial transaction, and improving working capital management.

But the most important indicator, according to Hadidi, is the extent to which the financial department benefits from the capabilities provided by automation to provide more useful analyzes and contribute to improving decisions, instead of being satisfied with completing traditional work faster.

The results of the “Financial Trends 2027” report issued by Deloitte indicate that 43 percent of participants placed the integration of artificial intelligence and advanced technologies for process automation among their top priorities until the fiscal year 2027.

In contrast, 34 percent identified enhancing enterprise-wide cost efficiencies, 33 percent improving capital allocation and investment decisions, and 32 percent identifying strengthening the strategic impact of financial management.

These numbers reflect the priorities of participants in the survey on which the report is based, and do not represent percentages specific to Saudi institutions.

From preparing numbers to interpreting them

Hadidi believes that one of the most prominent changes imposed by artificial intelligence on financial professions is the shift in the role of the employee from producing numbers and processing transactions to interpreting results, examining assumptions, and supporting decisions.

As repetitive work becomes automated, the need for specialists who are able to investigate unusual findings, assess the accuracy of analyses, and understand their implications for an organization's financial and business performance increases. This transformation is of particular importance in Saudi Arabia, as Hadidi points to the opportunity to develop national competencies in financial analyses, forecasting, and partnerships with business sectors, instead of relying on career paths that focus on implementing traditional accounting processes. This trend is linked to the objectives of the National Strategy for Data and Artificial Intelligence, which aims to enhance the Kingdom’s position in the field of artificial intelligence by 2030.