As the UAE accelerates its ambition to become a global artificial intelligence hub, concerns around data control, cybersecurity and governance are moving to the centre of discussions about how AI should be deployed.
Industry experts say the challenge is no longer simply where data is stored, but whether governments and organisations can maintain control over the infrastructure, governance and operations that support increasingly critical AI-driven services.
“Resilience ultimately depends on control, which is why the two have become inseparable,” said Mohammed Retmi, Vice President for Sovereign Clouds at Core42. “Sovereignty is about control: the ability to set your own risk tolerance for data and AI, and to keep that control during normal operations, disruption and recovery.”
The issue is gaining importance as governments around the world integrate AI into public services, financial systems and other critical sectors. In the UAE, where digital transformation is a national priority, policymakers have increasingly focused on ensuring that cloud and AI systems meet local security and regulatory requirements.
According to Retmi, digital sovereignty now extends beyond data residency to include operational and technological control. This means organisations must not only know where information is stored, but also who can access systems, how infrastructure is governed and whether services can continue operating securely during disruptions.
He argued that cybersecurity can no longer be treated as an afterthought. “Access controls, zero-trust architecture, encryption, continuous compliance monitoring and auditability must be engineered into the operating environment from the outset,” he said.
The UAE is already seeing AI deployed on a significant scale within government. Core42’s sovereign-enabled public cloud platform, developed with Microsoft for the Abu Dhabi Government through the Department of Government Enablement, supports more than 47 government entities and over 11 million digital interactions each day. The project aligns with Abu Dhabi’s goal of becoming the world’s first fully AI-native government by 2027.
As AI adoption expands, experts expect infrastructure requirements to become more complex. Training, fine-tuning and running AI applications each require different levels of computing power, speed and security.
“AI infrastructure itself will also become more heterogeneous,” Retmi said. “Organisations will therefore need greater flexibility to match each workload to the right infrastructure rather than being locked into a single architecture or technology roadmap.”
The debate is also extending to governance, particularly as organisations begin experimenting with autonomous AI agents capable of making decisions and carrying out tasks with limited human intervention.
“The shift from single-output AI to multi-agent systems changes the infrastructure requirement at every layer,” Retmi said. “Organisations need to know what each agent can access, how it behaves, how its actions are traced and who remains accountable for the outcome.”
Beyond technology, he warned that governance frameworks must address how employees use AI tools inside and outside approved environments. “The practical answer is to give people approved, governed AI inside the sovereign boundary,” he said.
As governments and businesses deepen their use of AI, industry leaders say success will increasingly depend on embedding governance and security into systems from the start, rather than trying to retrofit controls after deployment. “The organisations that scale AI successfully,” Retmi added, “will be those that build governance into deployment from the beginning.”
Source: Khaleej Times