The United Arab Emirates has set what is arguably the most ambitious public-sector AI target announced by any country to date. After a Cabinet meeting chaired by Sheikh Mohammed bin Rashid Al Maktoum, the UAE unveiled a framework to move 50% of federal government services, sectors, and operations to agentic AI within two years.
The announcement, made on April 23 and amplified across regional and international press over the past 48 hours, would make the UAE the first national government to operate at this scale through autonomous AI systems rather than human-driven workflows.
Sheikh Mohammed: "AI is no longer a tool"
"AI is no longer a tool. It analyses, decides, executes and improves in real time," Sheikh Mohammed said, framing autonomous models as an "executive partner" for accelerating decisions and raising service efficiency. The strategy is being directed by President Sheikh Mohamed, with Vice President and Deputy Prime Minister Sheikh Mansour bin Zayed overseeing implementation. Mohammed Al Gergawi, Minister of Cabinet Affairs, leads a dedicated task force.
The scope is unusually broad. The plan calls for AI agents — autonomous systems capable of making decisions and pursuing goals with limited supervision — to redesign policies, processes and procedures across federal entities. Every federal employee will receive mandatory AI training, dovetailing with a parallel Dubai-level program — Digital Dubai's AI Workforce Transformation Program (AI+) — that is reported to cover around 50,000 emirate-level government staff.
Building on existing digital infrastructure
The UAE is not starting from scratch. The agentic rollout is set to ride on top of existing platforms such as UAE Pass for digital identity and TAMM for citizen services, and it builds on a decade of policy moves: the appointment of the world's first AI minister in 2017, the founding of Mohamed bin Zayed University of Artificial Intelligence in 2019, and the introduction of AI as a school subject across all government education stages.
A phased introduction is planned. Federal departments will be assessed on adoption speed, implementation quality and AI mastery before broader rollout. Initial use cases are expected to focus on high-volume, low-complexity workloads such as automating case handling, supporting policy execution and operational decision-making.
Implementation reality and the agentic readiness question
Industry analysts cited in coverage of the announcement flagged execution as the main risk. "The real determinant of success will be agentic readiness at the data and process layer, not infrastructure," Manish Ranjan, software and cloud research director at IDC EMEA, told Computer Weekly. The UAE has spent heavily on compute, including a US-UAE collaboration on an Abu Dhabi AI campus with reported 5 gigawatts of data center capacity, but transferring decision authority to AI requires clean data pipelines, defined process boundaries, and clear escalation rules.
Why this matters
Most government AI programs to date have been narrow — chatbots for citizen queries, document classification, fraud screening. The UAE plan is qualitatively different because it places agentic systems inside the operational backbone of the state, with explicit human-in-the-loop frameworks defining which decisions can be fully automated. If the rollout lands close to its target, it will produce the first large national reference architecture for governing through autonomous agents — and a benchmark other governments will be compared against.
— Michael Ouroumis
