UAE and MBZUAI plan Agentic AI training for 80,000 civil servants

The UAE is quietly turning itself into a large-scale testbed for applied agentic AI in government. Training 80,000 federal employees — essentially the entire federal workforce — to design, supervise, and deploy agentic systems is less about flashy demos and more about institutional capacity-building. It signals that the country doesn’t just want to buy AI tools; it wants civil servants who understand workflows, risks, and governance well enough to embed AI into core public services.([edtechinnovationhub.com](https://www.edtechinnovationhub.com/news/uae-government-partners-with-mbzuai-to-train-80000-federal-staff-in-agentic-ai))

For the global race to AGI, that matters in two ways. First, it accelerates deployment: having a government this large systematically trained in agentic patterns will generate extensive real-world usage data, corner cases, and institutional knowledge that can feed back into model and platform design. Second, it’s a template other states can copy: a national AI strategy that couples flagship models with human-capital investment and explicit targets (50% of services running on agentic AI). This kind of program effectively assumes that something like AGI — or at least very capable agents — will be standard infrastructure in government and is working backward to get the people and processes ready.