Why critical infrastructure needs AI at the edge, not just intelligent systems

Across the Middle East, critical infrastructure is becoming intelligent at an unprecedented pace. Telecommunications networks are increasingly software-defined. Utility grids are filling up with connected metres, sensors and automated controls. Artificial intelligence is moving closer to where decisions are made, at the network edge.
This shift promises faster responses, lower latency and greater autonomy. But it also exposes a growing contradiction: intelligence at the edge is scaling faster than trust at the edge.
Fraud, settlement delays and data disputes are often framed as operational or compliance issues. In reality, they share a deeper root cause. They stem from the absence of a real-time, verifiable trust layer where events actually occur. AI can recognise patterns, but it cannot prove that an event truly happened. In sectors where money, identity and national infrastructure intersect, that distinction is decisive.

AI at the edge still depends on trust

Edge AI excels at acting quickly. It flags anomalies, scores risk and automates decisions close to the source of data. But it relies on one fragile assumption: that the data it consumes is authentic.
In telecommunications, this assumption fails routinely. SIM swap fraud, interconnect leakage and roaming disputes all originate at the edge, where activity happens, but verification is weak. In utilities, smart meters generate massive data volumes, yet enterprises and regulators still ask a basic question: can this data be independently proven accurate, untampered and attributable?
When AI is pushed to the edge without mechanisms to verify events at their source, it doesn’t reduce risk. It moves risk closer to the point of impact. Decisions become faster, but not necessarily more defensible.

Fraud and settlement are edge trust problems

Roaming settlement illustrates the issue clearly. Despite decades of standards, much of the industry still relies on delayed, file-based reconciliation. Disputes are resolved weeks or months after usage occurs. Cash flow is locked not because operators lack analytics, but because they lack shared, verifiable ground truth at the edge.
SIM-based fraud follows the same pattern. Identity compromise increasingly originates at the SIM itself, yet operators and enterprises cannot independently verify SIM lifecycle events in real time. Detection becomes reactive rather than preventative.
Utilities face an analogous challenge. Smart grids depend on sensor data to automate billing, enforce service-level agreements and detect losses. Without verifiable provenance at the edge, that data remains an operational input, not auditable evidence. These failures are not about tooling. They are failures of edge trust architecture.

The edge is becoming the system of record

As more data is generated and acted upon outside centralised data centres, the edge is quietly becoming the system of record. Decisions are made where events occur, at SIMs, gateways, meters and devices.
This requires a new model. Instead of reconciling data after the fact, infrastructure must be able to sign, timestamp and validate events as they happen, without exposing sensitive information. AI can then operate on evidence, not assumptions.
This is where Web3 technologies become relevant, not as speculative assets, but as coordination infrastructure. Cryptographic proofs, decentralised identifiers and shared consensus mechanisms allow multiple institutions to agree on what happened, when it happened, and who is accountable, without centralising control.
In our experience building Distli, the real shift comes from combining AI at the edge with verifiable event data. That combination moves systems from dispute resolution to dispute prevention.

Telecom and utilities are converging

What is striking is how similar the challenges are across sectors. Telecom operators want faster settlement and effective fraud prevention. Utility companies want trustworthy metering and auditable automation. Both operate regulated, mission-critical systems where decisions must be defensible, not just fast.
As infrastructure becomes more autonomous, trust at the edge becomes the bottleneck. Intelligence without proof creates speed without certainty. Proof without intelligence creates certainty without scale. The next generation of critical infrastructure will require both.
The future is not just AI at the edge. It is AI at the edge that can be trusted, where decisions are grounded in verifiable events and trust is built into the architecture rather than retrofitted through audits.
This article was contributed by Joel Curado (LinkedIn). He is the CEO of Distli and holds a PhD in Information Technology Sciences, specialising in blockchain systems for critical infrastructure. He has spent nearly two decades working at the intersection of telecommunications, AI and distributed systems.

Written By Innovation City Ecosystem.

This article was originally published on Khaleej Times.
You can find the original version here.