In an era when artificial intelligence is increasingly being used by criminals to create realistic voices, fake messages, and sophisticated fraud schemes, a British company decided to turn the same technology against scammers themselves.

The project is called Daisy. It was developed by British telecommunications company O2 together with specialists in artificial intelligence and cybersecurity.

The idea is simple but unusual: instead of simply blocking suspicious calls, the system intentionally keeps scammers occupied and takes away their most valuable resource — time.

When a fraudster calls a specially prepared number, the call is answered not by a real person, but by artificial intelligence with the voice of a friendly elderly woman.

However, Daisy is not like an ordinary voice assistant.

She was designed to behave like a real person who is not very familiar with technology and could easily become a victim of fraud.

During conversations, “Grandma Daisy” deliberately acts slowly and slightly confused:

• she takes a long time looking for her glasses;
• gets distracted by the kettle;
• tells stories about her family;
• talks about her cat;
• discusses knitting and everyday activities;
• mixes up numbers;
• asks for information to be repeated several times;
• provides fake bank card details;
• asks the same questions over and over again.

For a scammer, such a conversation becomes a real trap.

The fraudster believes they are speaking with a potential victim, so they continue trying to convince “Grandma” to transfer money, reveal a password, provide card details, or click on a dangerous link.

Some criminals spend 30-40 minutes talking with Daisy while trying to complete the scam.

But during all that time, they are occupied by artificial intelligence and cannot call real victims.

Essentially, Daisy turns the scammers’ own strategy against them: instead of a quick attack, they get a long and completely useless conversation.

According to O2, since its launch, the system has already “spoken” with more than a thousand phone scammers.

The main goal of the project is not only to protect users but also to study criminals’ methods, collect information about common fraud techniques, and make phone networks safer.

This has become especially relevant as scammers increasingly use new technologies:

• voice deepfakes;
• automated calls;
• personalized message generation;
• imitation of relatives’ voices.

Previously, artificial intelligence was often viewed as a tool that helped create new threats.

But Daisy demonstrates another side of the technology: AI can not only help people, it can also become a “digital defender” itself.

Perhaps in the future, similar AI agents will automatically answer suspicious calls, keep scammers engaged in long conversations, and protect millions of people from financial losses.

For the first time in history, phone scammers have encountered an opponent that can do exactly what they do — persuade, talk, and waste time.

With one major difference: Daisy never gets tired, never gets nervous, and never transfers money.