THE United Arab Emirates plans to move half of its federal government sectors, services and operations to agentic AI within two years. The initiative is intended to create systems that can take proactive action, improve public services and reduce operating costs.
At a World Economic Forum event in Dubai on 14 October 2026, Minister of State for Artificial Intelligence Omar Sultan Al Olama said the government was using a “control tower agent” and continuous audits by both AI agents and people to oversee activity and reduce the risk of harm to citizens.
The project is being rolled out in phases. The first focuses on building government capabilities, assessing services and operations, and setting priorities; a later phase will establish FedAI, a national technical ecosystem for government agencies to develop and run agentic services. The government also plans to train staff in agentic AI and generative AI.
The article gives no technical detail about how the control tower works, but AmanTrue chief executive Evin McMullen said institutions should be able to identify which agent acted, who authorised it, and the task’s defined limits.
The article cites a ServiceNow report published on 20 October saying 29% of security teams had found unregistered AI agents running in production, while nine in 10 organisations had experienced a security incident attributed to agentic AI in the previous year. These figures are not evidence of incidents in UAE government systems. Al Olama warned that a harmful incident could undermine public trust in the technology.