OpenAI Bets AED 13 Billion That Rules Can Be a Product

AED 13 billion is the stake, and the bet is that regulation itself can become a commercial-grade technology asset. Abu Dhabi’s government and OpenAI have launched Machine-Readable Government, a framework that converts selected policies and regulations into machine-readable rules, letting AI systems assess routine cases and return decisions linked directly to the clauses behind them. The economics are simple: routine approvals that once took days of manual review could be processed within minutes, around the clock, freeing administrative capacity while speeding up the permits and licences that businesses depend on.

The programme sits inside the Abu Dhabi Government Digital Strategy 2025-2027, the emirate’s AED 13 billion digital commitment, and will integrate with TAMM, the platform that brings together more than 1,200 public and private services. Use cases are already active across 40 government entities. Planned applications include construction permits, business and freelance licensing, data access governance, HR compliance and procurement, a pipeline that covers much of the approval infrastructure shaping project timelines and the cost of doing business.

Governance and financial responsibility rest with the Department of Government Enablement Abu Dhabi, which leads the initiative. Officials keep final decisions, and human judgement is retained for complex cases. The framework is not intended to make autonomous decisions in high-stakes areas, nor in cases materially affecting individual rights, eligibility or access to essential services, a design constraint that limits automation’s scope while keeping accountability with the public sector.

The first pilot is already live. It runs with the Abu Dhabi Agriculture and Food Safety Authority for food imports, in collaboration with G42 company Inception42, converting written policy into automated compliance checks and reducing the manual burden on trade-critical processes. Beyond operational savings, structuring regulations in a common format carries a broader return: officials can identify overlaps or conflicts between departments and test proposed rule changes before introducing them, lowering the regulatory risk that investors and market entrants typically price in.

Meanwhile, the compute side reflects a growing market for locally hosted AI capacity. Processing runs through OpenAI’s UAE inference residency, on GPUs in UAE data centres under local cybersecurity requirements. Abu Dhabi Government retains ownership of the data generated through the system, and that data is not used to train AI models. The structure hands the government operator control of a strategic asset while anchoring processing activity inside the country’s data-centre market.

Read as a capital story, the initiative positions a government as both owner of digitised regulatory infrastructure and customer of commercial AI capability. OpenAI supplies the underlying models; G42’s Inception42 supports the pilot. The Department of Government Enablement acts as governing operator, setting automation’s boundaries and holding final decision authority. For private companies and entrepreneurs navigating licensing, procurement and permits through TAMM, the promised effect is faster turnaround at lower administrative cost, with decisions traceable to the exact regulatory clauses that produced them.

The strategy’s success will rest on whether machine-readable rules can extend across the 40 entities already working with the framework and integrate smoothly into the 1,200-service TAMM platform without eroding the human oversight officials have committed to preserve. If it does, Abu Dhabi will have converted its AED 13 billion digital investment into a regulatory operating system that compresses approval times from days to minutes, keeping data ownership, and the final word on every case, inside government hands.