Experts warned that telecom operators around the world are introducing AI to manage their autonomous network, where human oversight is essential to ensure safety, regulatory compliance, and ethical accountability in rapidly evolving systems.
As AI agents in telecom transform the world’s industry, alarms are being raised by experts: the increasing autonomy of AI needs to be monitored more closely. Whether through real-time diagnostics or predictive maintenance of the network, telecom operators are integrating sophisticated AI tools—many driven by multimodal and agentic systems—into the core of their infrastructure.
But as telcos’ autonomous system in networking has the potential to deliver speed and efficiency, risks from unregulated automation are increasing as well.
Autonomous Telecom AI systems can now process text, video, and sensor data in real time, offering insights and taking actions with little human intervention, but the transition comes with concerns. Over time, machine learning models evolve, often reacting to new inputs in unpredictable ways, prompting operators to set up performance dashboards and redeploying analytics teams to monitor data drift and model accuracy.
“Nothing is static,” noted CEO of TUPL Petri Hautakagas, whose ML model expanded from six to 40 data feeds. His point underscores the dynamic nature of telecom environments and the need for continuous human validation.
From Support Systems to Self-Acting Agents
The industry is now shifting from using AI network automation as a decision-support tool to relying on fully autonomous agents that act without human prompts. Agentic AI enables systems not just to recommend actions but to execute them, raising new questions about control and responsibility.
“Agentic AI will give customers more power to execute steps and functions – to get an answer, and choose an action without direct human involvement,” said Steve Szabo of Verizon Business.
Meanwhile, supporting an autonomous network of this transformation is also changing where we see multi-modal models and mixture-of-experts (MoE) architectures are now capable of merging varying forms of data—weather forecast to building activity—to anticipate network outages, such as signal degradation caused by cranes or traffic congestion.
Spirent’s Stephen Douglas said, “The enterprise doesn’t want a black box,” and businesses must create AI systems that are transparent and explainable to users and regulators alike.
With telecoms moving deeper into autonomous network AI, it will be innovation balanced with regulation that will be key to avoiding disruption, maintaining public trust, and ensuring ethical AI integration in a fast-digitalizing world.
Inside Telecom provides you with an extensive list of content covering all aspects of the tech industry. Keep an eye on our Telecom sections to stay informed and up-to-date with our daily articles.