AI-Assisted Data Cleansing — Jordan Restores 2 Million Tawjihi Exam Records and Harmonises 2.5 Million Land Records

Jordan's Ministry of Digital Economy and Entrepreneurship signed an agreement with the Japan International Cooperation Agency (JICA) in December 2024 to apply AI to two long-standing government data-quality problems, with implementation carried out by Amman-based firm Eqra Tech under the project 'Using Artificial Intelligence for Government Data Cleansing in Selected Use Cases.'

The project aimed to repair and harmonise two large, long-standing government datasets that conventional methods had failed to fix, as a foundation for more reliable analytics and decision-making rather than a citizen-facing service in its own right.

The first use case restored national-ID linkages on more than 2 million Tawjihi (General Secondary Education Certificate) exam records spanning 1985-2004, after conventional data-recovery methods had failed to reconnect historical results to citizens' national ID numbers. The second harmonised more than 2.5 million land records shared between the Ministry of Local Administration and the Department of Lands and Survey, unifying inconsistent coding systems between the two agencies' databases.

Both projects were reported completed by mid-2025, restoring over 2 million exam records and harmonising over 2.5 million land records.

No source found discloses a numeric accuracy or match-rate figure, an independent audit, or a downstream measure of how the cleaned data has since improved a specific government service. The project is one of the more concretely completed items in Jordan's 2023-2027 AI Strategy and Implementation Plan, a 68-project roadmap on which most other announced activity remains at the strategy or training stage.