Tariq Saeedi
When high-level security summits assemble to confront the rise of Artificial Intelligence, the narrative almost invariably follows a familiar script. Speakers warn of malicious actors weaponizing new tools—terrorist networks generating automated propaganda, extremist groups deploying hyper-realistic deepfakes, or bad actors using generative algorithms to streamline recruitment across digital borders.
In this conventional framework, the threat model is straightforward: a bad actor takes a conscious technological tool, points it at a target, and pulls the trigger. AI is viewed strictly as a weapon in human hands.
Yet, focusing exclusively on intentional misuse misses a far broader and more fundamental transformation. The true challenge of the modern era is not merely how AI might be targeted against security, but how AI has quietly become the very infrastructure upon which our physical, financial, and administrative systems rest.
When automation is woven into the fabric of daily life, the boundary between general security and AI security vanishes entirely. The two concepts superimpose onto one another until every conventional security scenario becomes, at its core, an AI security scenario.
To understand this shift, we have to look past conscious schemes and examine two critical structural thresholds: the entry point of unwitting exposure, and the exit point of cascading risk.
The Entry Point: Innovation Without Insulation
The entry point into this dilemma occurs when AI is introduced into critical environments—energy networks, port logistics, municipal traffic grids, banking clearinghouses, or public administration—simply for the sake of efficiency, speed, or optimization, but without rigorous, built-in security considerations.
In these cases, those who introduce AI are not scheming against security; they are trying to solve routine operational problems. However, by embedding autonomous systems into legacy environments without adequate oversight, a latent vulnerability is created.
The threat actor in this scenario does not need to master advanced AI techniques or write sophisticated algorithms. They can rely on traditional, low-tech intrusion methods—manipulating a sensor, corrupting a simple data feed, or severing a physical communication link.
But because an autonomous AI engine now governs the underlying process, that small, ordinary intrusion triggers an exaggerated, automated reaction. The vulnerability wasn’t created by an AI attack; it was created by unwittingly handing control to an algorithm before securing its inputs.
The Exit Point: The Unintended Cascade
The exit point is where this structural vulnerability manifests as a real-world security crisis.
Imagine a traditional physical disturbance: an unexpected localized blackout, a supply chain bottleneck at a border crossing, or a coordinated misdirection campaign using basic low-tech channels. In a legacy system, human operators would evaluate the context, exercise discretion, and absorb the shock.
In an AI-integrated environment, automated control systems react according to their strict optimization training. Operating at machine speed without human intuition, the system may execute logical but catastrophic decisions—shutting down connected sub-grids, re-routing essential resources away from critical areas, or flagging thousands of false positives that overwhelm security personnel.
Here, AI was neither the architect nor the conscious weapon of the threat. Yet, its presence acted as an amplifier, transforming a minor physical risk into a systemic security failure.
Shifting the Governance Paradigm
If we continue to view technology solely through the lens of deliberate misuse, our policy responses will remain reactive and narrow—focused on taking down bad content, blocking specific apps, or chasing individual bad actors.
A complete view of systemic risk demands that state leaders, intelligence bodies, and tech developers adopt a policy of Security-by-Design. As governments across Central Asia and the broader global community work to establish regional cyber hubs, AI governance frameworks, and international codes of conduct, the emphasis must extend beyond preventing digital propaganda.
We must begin asking the hard structural questions: Where have we introduced automated decision-making without sufficient safeguards? What happens to our essential infrastructure when an algorithm faces an unscripted physical crisis?
AI is no longer just a tool sitting on the security workbench; it is the floor upon which the workbench stands. Ensuring long-term stability requires making sure that floor is solid before we place the weight of global governance upon it. /// nCa, 9 November 2026