As governments and businesses race to adopt artificial intelligence at scale, a central question is emerging: how do humans and AI agents work together without compromising trust, safety, or innovation?
That question took center stage at the World Government Summit’s AI Forum, where global technology leaders discussed how autonomous AI agents are reshaping productivity and why human oversight, governance, and infrastructure must evolve alongside the technology. The session brought together Simon de Montfort Walker, Executive Vice President of Industry Applications at Oracle, Masood Mohamed Sharif, CEO of e& UAE, and Carl Pei, Co-founder and CEO of Nothing.
Panelists agreed that the next phase of AI is no longer about isolated tools but about teams of intelligent agents collaborating with humans. Walker described this shift as the evolution of automation, moving clerical and repetitive tasks away from people while placing humans at the center of decision-making and oversight. “The goal is to remove the clerical work and let humans manage outcomes,” he said, noting that organisations are increasingly deploying specialised AI agents with defined roles, working together like a digital assembly line. He also emphasized that organisations investing early in organising and exposing their data are gaining a lead in real-world AI adoption.
For Sharif, the rise of AI agents carries particular implications for telecom operators, positioned at the intersection of connectivity, compute power, and data. As e& transforms from a traditional telecom company into a “techco,” he said AI impacts infrastructure enablement, internal operations, and customer relationships. AI requires ultra-low-latency connectivity, massive computing capacity, and efficient data routing, making telecom providers central to building an “AI fabric.” e& has already deployed hundreds of AI use cases, from back-office automation to visual fault detection in home routers, reducing truck rolls by up to 40 per cent in some instances. Sharif noted that most large enterprises will not develop AI models from scratch but will rely on partnerships with global technology companies and agile startups.
Pei highlighted how AI agents are reshaping software development. Recent AI coding tools now allow even non-technical users to build fully functional applications, collapsing the traditional software stack. He predicted a future in which individuals, aided by AI, can develop products that once required entire teams, with “always-on” AI agents continuously building, testing, and refining software around the clock.
The discussion also addressed governance and ethics. Sharif stressed that sovereignty involves cybersecurity, data privacy, ethical standards, and human oversight, not just cloud ownership. He cautioned that over-regulation could stifle innovation while highlighting the importance of continuous monitoring and human training to prevent misuse of AI systems.
Speakers pointed to the UAE as a model for balancing experimentation with policy and infrastructure. Walker praised the country’s ability to combine technical execution with policy ambition, while Sharif called AI a “beautiful disruption” that will create new sectors, boost productivity, and expand opportunities globally.
