AI readiness in official statistics: from regional cooperation to self-assessment

The UN Statistics Division (UNSD) of UN DESA will hold a capacity-development event called 'Artificial Intelligence and Official Statistics: Strengthening Capacities and Readiness' on Sunday 8 November 2026, from 09:00 to 12:55. It takes place alongside the 2026 UN World Data Forum, which runs in Riyadh from 9 to 12 November 2026. Only registered attendees of the Forum can enter the venue. No registration deadline is given, and a form is available for programme updates.

The event aims to help participants assess what AI means for the production and use of official statistics. It covers two areas: using AI in statistical production, and making official statistics usable by AI. Participants are expected to assess the benefits and limitations of AI based on national experience. They will also identify improvements to data publication that support reliable AI access and reuse, and describe their institutional capacity needs.

The programme opens at 09:00. Session 1 starts at 09:10 and covers coding, classification, processing and validation. The speakers are Rohit Bhardwaj (MoSPI, India), Nikola Orlic (Statistical Office of the Republic of Serbia), Eshan Romjon (Statistics Mauritius) and Isabella Schmidt (UN Women). Session 2 starts at 10:10 and covers making official statistics usable by AI.

The 17th session of the Statistical Committee of the UN Economic and Social Commission for Western Asia (ESCWA) ended on 9 September 2026 in Jabal Akhdar, Oman. Oman's National Centre for Statistics and Information (NCSI) hosted the session, which brought together the heads of Arab statistical offices.

The Committee adopted a set of recommendations. One of them is to expand the use of modern technology, artificial intelligence and non-traditional data sources in producing official statistics. The most notable recommendation is to establish the 'Muscat Team' on using AI in household income and expenditure surveys. The team would build on Oman's experience of running that survey with AI.

NCSI CEO Dr. Khalifa Abdullah Al Barwani said the team will be a framework for Arab statistical cooperation and for exchanging AI expertise in official statistics. These outcomes are recommendations of the Committee.

Oman's National Centre for Statistics and Information (NCSI) launched the results of its project 'Household Income and Expenditure Survey using artificial intelligence' on the evening of 6 September 2026. The Oman News Agency reported that the launch ceremony took place at NCSI headquarters in Muscat. NCSI describes the project as a national achievement and the first of its kind worldwide. This is NCSI's own claim.

NCSI CEO Dr. Khalifa bin Abdullah Al Barwani said that using AI makes the production of statistical indicators more efficient and sustainable. He said it needs less time, effort and resources than traditional methods while meeting the highest requirements of quality, accuracy and reliability. According to him, the survey provides integrated data on the size and patterns of household consumer spending and how they relate to demographic variables. It also provides the expenditure weights needed to compile the Consumer Price Index, and data on living conditions that meets the requirements of the System of National Accounts.

He added that national staff delivered the project through institutional integration and that it is the start of a wider use of AI. The source does not describe which AI methods were used.

The World Bank Group offers an AI-Readiness Assessment Framework for National Statistical Organizations. The Development Data Group and the Office of the WBG Chief Statistician developed it, and it is released under a CC BY 4.0 licence. The framework has two pillars. 'AI for Data' asks whether an NSO can use AI responsibly. 'Data for AI' asks whether AI can use official statistics reliably.

The framework has 12 dimensions, 69 questions plus a gateway question, and four maturity levels from A to D. Offices can use it on their own or in a facilitated assessment. A web form scores the answers and produces a readiness profile with priority gaps and actions in a suggested order.

The questions are linked to the recommendations in the CCSA background paper 'AI-Ready Official Statistics'. That paper was prepared for the 57th session of the UN Statistical Commission in March 2026 and contains 39 recommendations for NSOs. Central banks and ministries that produce official statistics can also use the framework.

Saudi Arabia's General Authority for Statistics (GASTAT) reports that 33.1% of establishments used artificial intelligence technologies in 2025. The figure comes from Table 16 of its Establishments ICT Access and Usage Statistics 2025, which shows AI use for 18 economic activities.

The highest share was in information and communication (61.1%). It was followed by financial and insurance activities (52.9%), education (51%), transportation and storage (44.4%), professional activities (43.9%), and mining and quarrying (41.8%). The lowest shares were in other services (26.7%), wholesale and retail trade (30.2%), administrative and support services (30.4%), and manufacturing (30.7%).

This table shows how an NSO can measure AI adoption in the economy with its existing establishment ICT survey and break the results down by sector.