Abu Dhabi has officially launched the world’s first live deployment of Machine-Readable Government, an ambitious initiative built in partnership with OpenAI designed to streamline public services through artificial intelligence. In its initial operational phase, the framework is being deployed at the Abu Dhabi Agriculture and Food Safety Authority (ADAFSA) to revolutionize food import checks. Under the new system, each regulatory case is automatically cross-referenced against a defined set of digitized laws and policies, with every outcome directly traceable to the specific legal clause that governed it.
The initiative represents a fundamental shift in how public administration operates, initially focusing on transforming the application of existing rules rather than altering underlying legislation. Crucially, human oversight remains firmly embedded within the architecture; government officials retain ultimate responsibility for final decision-making, ensuring accountability and compliance with national standards.
The integration of this technology marks a major milestone in a broader series of agreements between Abu Dhabi and OpenAI. The ADAFSA pilot has moved swiftly into production, tackling a critical sector for the United Arab Emirates, which currently imports approximately 90 percent of its food supply. Alongside the deployment of advanced AI tools, the UAE government is actively pursuing broader strategies to reduce this heavy reliance on foreign markets by expanding domestic processing capacity, reinforcing a national push toward long-term food self-reliance and supply chain security.
The pioneering pilot was developed in collaboration with Inception42, a prominent G42 company, utilizing a specialized Policy Adoption Engine designed to translate complex written government policies into automated compliance checks. In its initial rollout, Machine-Readable Government is supporting routine regulatory assessments, including applications for various permits and certificates. Every evaluation generates a clear audit trail by linking the final assessment directly to the statutory clause behind the outcome.
Despite the rapid deployment, several operational details remain open for importers navigating the new digital landscape. Publicly available reports have yet to specify the exact subset of food import regulations currently monitored by the system, the precise volume of applications handled by ADAFSA since implementation, or the average processing speeds achieved. It also remains unclear whether halal certificate verifications are directly incorporated into the automated checks, how applicants are formally notified when a decision has been assisted by artificial intelligence, how the underlying machine-readable rules are dynamically updated when national regulations evolve, or the total financial value of the contract. Overall governance and leadership of the initiative are spearheaded by Abu Dhabi’s Department of Government Enablement (DGE).
Officials Keep Final Decisions as AI Handles Routine Cases
Under the Machine-Readable Government framework, human oversight remains a core pillar of public administration. Officials are responsible for setting regulatory standards, reviewing complex or nuanced applications, and maintaining ultimate authority over final decisions. Agency leadership has emphasized that the framework is explicitly not designed for autonomous decision-making in high-stakes sectors, nor in sensitive cases that could materially affect individual rights, eligibility, or access to essential public services.
By automating routine administrative tasks, the system is designed to process standard applications autonomously around the clock. This capability is expected to dramatically accelerate bureaucratic workflows, potentially compressing processes that previously required days of manual review down to mere minutes. In addition to handling standard transactions, the technology provides officials with advanced analytical capabilities, allowing them to detect potential overlaps or conflicts between regulations across different government departments and test proposed policy changes against existing legal frameworks before implementation.
Ahmed Tamim Hisham Al Kuttab, chairman of the DGE, described the framework as the vital connective tissue linking Abu Dhabi’s overarching artificial intelligence ambitions with practical, everyday implementation. Meanwhile, Sam Altman, chief executive officer of OpenAI, highlighted the significance of the milestone, noting that by transforming specific laws into a machine-readable format while keeping human decision-makers firmly in control, Abu Dhabi has established a new global standard for AI-enabled public services.
Machine-Readable Government Runs AI Processing Inside the UAE
To address critical data security and sovereignty requirements, all artificial intelligence processing for the Machine-Readable Government initiative is executed strictly within national infrastructure inside the UAE. Under OpenAI’s dedicated UAE inference residency model, real-time computation runs on advanced graphics processing units housed within domestic data centers. The platform operates on a secure "fail-closed" architecture governed by stringent national cybersecurity standards and fully subject to Abu Dhabi law.
Government authorities maintain absolute ownership over all data generated through the system. Crucially, this operational data is strictly quarantined and will not be utilized to train external AI models. Comprehensive safeguards have been established to regulate how information is securely stored, utilized, and eventually deleted. Industry observers note that these measures position the UAE as the first country in the Middle East to leverage OpenAI’s advanced GPT models through in-country inference infrastructure.
ADAFSA Moved Food Import Services Online Years Before MRG
The introduction of Machine-Readable Government builds upon a foundation of prior digital transformation within the agriculture and food safety sector. ADAFSA previously developed significant digital import infrastructure years prior to the OpenAI partnership, though these legacy systems operate independently of the new AI framework. In October 2022, ADAFSA launched the Food Import and Export Management Information System at GITEX Technology Week, deployed via the Advanced Trade and Logistics Platform operated by Abu Dhabi Ports’ Maqta Gateway.
Designed as an integrated single-window portal for importers and exporters, the platform enabled stakeholders to release food shipments prior to their physical arrival at port while tracking import and export orders, customs clearance, and pre-shipment inspections. That digital rollout successfully brought seven core ADAFSA services online, including corporate registrations for food importers, imported food product approvals, import and export certificate requests, and customs declarations.
Earlier in 2022, ADAFSA also introduced a smart analytics system utilizing artificial intelligence and machine learning to forecast import quantities and costs for agricultural, animal, and poultry products. Built on historical trade data spanning from 2015 to 2020, that system targeted key commodities such as beef, eggs, frozen chicken, and tomatoes to help identify new supply chain diversification opportunities and minimize waste across the domestic market.
Construction Permits and Business Licensing Follow Food Imports
Following the initial deployment in food import checks, the Machine-Readable Government framework is slated for expansion into several other vital public sectors. Upcoming applications include the automation of construction permits, business and freelance licensing, data access governance, human resources compliance, and government procurement. Furthermore, the MRG initiative will integrate seamlessly with TAMM, Abu Dhabi’s unified government services portal, which currently consolidates more than 1,200 public and private services. Across the wider public sector, artificial intelligence use cases are already active across 40 distinct government entities.
The rapid scaling of these technologies follows high-level strategic meetings, including discussions between Sheikh Khaled bin Mohamed, Crown Prince of Abu Dhabi and Chairman of the Abu Dhabi Executive Council, and OpenAI CEO Sam Altman, which resulted in a series of major bilateral agreements. Among these accords is a strategic partnership involving Abu Dhabi’s Mubadala investment vehicle and Inception42 to collaborate on AI-driven investment analysis, opportunity identification, and risk assessment. Additionally, a five-year agreement was established for Inception42 to adopt OpenAI’s Frontier platform across government administration, enterprise, and educational institutions.
The Machine-Readable Government initiative serves as a cornerstone of Abu Dhabi’s sweeping Dh13 billion Government Digital Strategy spanning from 2025 to 2027. Led by the DGE, the strategy aims to achieve total automation across government administrative processes and is projected to contribute more than Dh24 billion to the local economy by 2027, marking a pivotal step toward Abu Dhabi’s goal of establishing the world’s first fully AI-native government.