Operators are testing AI control loops in 5G slicing
AI News reports Nokia and AWS are piloting an agentic setup where network resources can be adjusted in response to live conditions. The tests involve du in the UAE and Orange in Europe and Africa.
The core idea is simple. Network slicing already exists in 5G, but many decisions are still manually tuned. Nokia’s framing is that AI agents can close that gap by adapting slices to changing demand with less human intervention.
This is less about hype and more about operations
The important question is whether operators can move from demos to repeatable control. Telecom systems carry critical traffic, so automation is only useful if reliability and accountability improve at the same time.
AWS appears as the model and platform layer, while Nokia appears as the telecom control layer. If that split works in production, it may become a common architecture pattern for operators that want cloud-scale AI without rewriting core network workflows.
For teams evaluating architecture options, keeping a central AI models reference helps align model behavior with network constraints.