O2’s New AI Granny Takes on Phone Scammers With a Dose of Sass and Strategy

O2 has launched "dAIsy," an AI-powered chatbot that impersonates a chatty grandmother to waste the time of phone scammers. Using voice transcription, a custom language model, and text-to-speech technology, dAIsy keeps fraudsters engaged with off-topic banter. This creative tool not only disrupts scam attempts but also highlights how AI can play a proactive role in protecting vulnerable individuals from fraud.

USAGETOOLS

The AI Maker

5/5/20252 min read

AI Granny answering questions on a phone
AI Granny answering questions on a phone

In a creative twist on combating digital fraud, the U.K.’s largest mobile network, O2, has unveiled a chatbot with a unique mission — waste the time of phone scammers until they give up. Named “dAIsy,” this artificial intelligence tool poses as a friendly, long-winded grandmother who’s eager to talk about everything but handing over her bank details. Her weapons? Small talk about knitting, her beloved cat Fluffy, and the kind of leisurely confusion that can drive even the most persistent scammer to hang up.

At a time when phone scams are becoming more sophisticated and harder to detect — especially with the rise of voice cloning — dAIsy offers a surprising but clever counterattack: using AI not just to detect fraud, but to actively engage and disrupt the fraudsters themselves.

Here’s how it works: dAIsy is powered by a combination of AI technologies. When a scammer calls, their voice is transcribed into text. A custom language model then crafts a response that fits dAIsy’s quirky grandmother persona. That reply is passed through a text-to-speech engine, which gives it voice — a convincingly warm, older-sounding tone, complete with the occasional ramble or off-topic comment. The result is a fully interactive, time-wasting decoy that scammers might think is a real target.

The system was developed in collaboration with Jim Browning, a well-known “scambaiter” who has spent years recording and exposing scam operations on his popular YouTube channel. His insights into the psychology and tactics of fraudsters helped train the AI to respond in ways that keep the caller on the line longer — a tactic that not only frustrates the scammer but potentially prevents them from targeting someone else.

The idea is not just amusing — it’s needed. According to FBI reports, Americans over 60 lost a staggering $3.4 billion to phone scams in 2023, a sharp increase from the year before. These scams often rely on urgency and emotional manipulation, targeting older adults who may be less tech-savvy or more trusting over the phone. With the growing sophistication of AI-generated voices, even more people are at risk of being duped.

That’s what makes dAIsy such a promising approach. Rather than trying to block every scam — a nearly impossible task — it flips the script by wasting the scammers’ time, reducing their efficiency, and collecting data that could be used to improve future protections.

O2’s initiative shows how AI can be used not just to automate or analyze, but to play defense in a creative, human-centered way. And if that defense happens to involve cat stories and the occasional confusion about whether it’s Tuesday or Thursday — even better.

As phone scams continue to evolve, so must the tools we use to stop them. dAIsy is more than just a novelty — she’s a knitting, chatting, time-sucking glimpse into the future of scam prevention.

Cited: https://techcrunch.com/2024/11/15/ai-granny-is-happy-to-talk-with-phone-scammers-all-day/