AI to block scam calls before you pick up the phone

 ChatGPT
Фото: ChatGPT

A new mechanism based on artificial intelligence is designed to to block scam calls even before you pick up the phone, El.kz reports.

How it works

During a government meeting on April 28, Deputy PM and Minister of Artificial Intelligence and Digital Development Zhaslan Madiyev explained the logic behind the new system. Mobile operators will begin using AI not only to improve network quality but also to recognize fraudulent schemes in real time.

"In cases of call spoofing, or signs of a fraudulent scheme, the system blocks the call in a fraction of a second even before the subscriber picks up," Zhaslan Madiyev said.

This is fundamentally different from how the fight against telephone fraud worked before which relied on complaints and manual number blocking.

Not just protection - a smart network too

In parallel, the same system will monitor the state of the telecommunications infrastructure. According to Madiev, the AI analyzes and predicts network load, user density, and the placement of base stations.

"This will help automatically redistribute resources without the involvement of engineers," the minister said.

This means the network will be able to adapt to real-time loads automatically, without manual intervention from specialists.

Why it matters now

The announcement comes amid fresh data on phone fraud in Kazakhstan. In April alone, the National Security Committee (KNB) in Almaty thwarted a scheme through which attackers made more than 4 million calls to Kazakhstanis.

According to law enforcement, the scammers used virtual stations, local numbers, and malware to bypass operators and disguise their calls as legitimate. The damage to telecom operators exceeded 15 million tenge.

Against this backdrop, the AI filter is seen no longer as an experiment, but as an attempt to close vulnerabilities at the network level itself. If the system works as described by the ministry, scammers will have far fewer chances of reaching subscribers.

El recommends