OpenAI and Abu Dhabi launch machine-readable government platform
Abu Dhabi's Department of Government Enablement (DGE) and OpenAI have jointly launched Machine-Readable Government (MRG), a platform that encodes a defined set of emirate laws and regulations into structured, machine-readable formats so that AI systems can assist officials with routine compliance decisions. The initiative forms part of Abu Dhabi's stated aim to become the world's first AI-native government by 2027 and is backed by the AED 13 billion ($3.5 billion) Abu Dhabi Government Digital Strategy running from 2025 to 2027.
The first production use case is live at the Abu Dhabi Agriculture and Food Safety Authority (ADAFSA), where the system automates food-import compliance checks. Each decision generated by the platform is linked to the precise regulatory clause that produced it, creating an audit trail that officials can inspect. Routine cases that previously required days of manual review can now return decisions within minutes, the government says. Human oversight is explicitly retained: government teams set standards, review complex cases, and remain accountable for every final outcome. The system is not intended to make autonomous determinations in areas that materially affect individuals' rights or access to essential services.
Infrastructure and data sovereignty
AI inference for MRG runs on OpenAI's UAE inference residency, meaning model computation is confined to GPUs located within UAE data centres. The architecture operates on a fail-closed security model governed by UAE cybersecurity standards and Abu Dhabi law. The Abu Dhabi government retains full ownership of all data generated by the system; that data is not used to train OpenAI models, and jointly agreed rules govern retention, use and deletion. The release describes this as the first deployment of OpenAI's GPT models using in-country inference infrastructure in the Middle East.
The technical implementation is built around a reusable Policy Adoption Engine, developed in collaboration with Inception42, a subsidiary of the Abu Dhabi-based conglomerate G42. The engine converts narrative policy text into automated compliance checks. Integration with TAMM, the emirate's unified service application consolidating more than 1,200 public and private services, is planned, with upcoming applications covering construction permitting, business and freelance licensing, data access governance, HR compliance, and procurement.
Market and regulatory context
The MRG initiative lands at a moment when governments worldwide are debating how AI can be embedded into public administration, and on what terms. The EU AI Act classifies AI systems used in public administration for decisions affecting individuals' access to services as high-risk, requiring conformity assessments, human oversight provisions, and transparency obligations. Abu Dhabi is not subject to EU rules, but the architecture's design choices, traceable decisions, retained human accountability, and data-residency controls, broadly mirror the safeguards that high-risk classification demands under European frameworks. That alignment may matter as the emirate seeks to attract international businesses and investors who must satisfy their own home-country regulators.
The broader market for govtech AI is contested. Palantir has long-standing contracts with Western defence and intelligence agencies for decision-support tooling; Microsoft and Google are competing for government cloud and AI contracts across Europe, the Gulf, and Asia-Pacific. OpenAI's direct government engagement, including its prior partnership with the US Department of Defense and now this Abu Dhabi initiative, signals an increasingly assertive push into sovereign and public-sector channels.
The machine-readable policy concept itself, sometimes called policy-as-code, has been explored in academic and standards contexts for years but has rarely reached production at a whole-of-government scale. Whether Abu Dhabi's rollout demonstrates a replicable model for other administrations will depend heavily on how the platform performs as scope expands beyond a single regulatory body. Named milestones to watch include the construction permitting and business licensing rollouts, and the point at which MRG integrates fully with TAMM's service catalogue.