Cisco has warned that the next wave of cyber threats will target the AI supply chain and the autonomous decision-making processes of AI agents as enterprises across the Middle East accelerate adoption.
According to the Cisco AI Readiness Index 2025, 92% of organisations in the UAE and 91% in Saudi Arabia plan to deploy AI agents across sectors including government, finance, and energy. These systems are designed to execute complex tasks without human intervention, shifting AI from experimentation into operational reality.
Cisco cautions that this rapid rollout is outpacing traditional security models. Unlike conventional software, AI systems depend on a complex ecosystem of third-party model files, open-source datasets, and specialised infrastructure. A single compromised component in this “AI supply chain” can expose enterprises to arbitrary code execution or large-scale data exfiltration.
“As AI agents move from experimentation to real-world deployment, organizations are facing new security considerations,” said Fady Younes, Managing Director for Cybersecurity at Cisco Middle East, Türkiye, Africa, and Romania. “From the third-party components used to build AI systems to how autonomous agents interact with data and tools, securing the full AI lifecycle is becoming essential for digital trust.”
The rise of agentic AI introduces more severe attack vectors. Early generative AI risks centred on prompt injection, where models were tricked into producing harmful outputs. AI agents, however, can act. If compromised, they could delete databases, send fraudulent communications, or manipulate connected enterprise tools.
To address this, Cisco has enhanced its AI Defense platform to function as a secure gateway. The technology intercepts interactions between AI agents and enterprise systems, identifying malicious instructions before they are executed.
The urgency reflects broader national ambitions. With over 74% of UAE organisations ranking AI as a top IT priority, focus is shifting from building AI to securing it. Cisco’s regional roadmap emphasises vulnerability scanning for third-party AI components, real-time runtime guardrails, and zero-trust identity controls for AI agents.
As the Middle East positions itself as a global AI hub, the ability to secure autonomous systems will shape the region’s digital future.
