Telephone scams cost elderly victims billions each year, causing financial loss and emotional distress. Now, AI voice agents — trained to sound just like vulnerable older people — are turning the tables.
These “AI grandmas” keep scammers talking for minutes, even hours, while filtering genuine calls straight to their real recipients. It’s a clever mix of humour and high-tech defence that’s protecting some of society’s most at-risk individuals.
Why AI Voice Agents Are a Scammer’s Worst Nightmare
Scammers rely on quick, easy access to trusting targets. They look for people who are polite, a little confused, and unlikely to hang up. AI voice agents can imitate exactly that — but with a twist: they never hand over personal details.
Instead, they ask for repetition, wander off into unrelated stories, and draw out the conversation until the scammer realises they’ve been wasting their own time.
Real-World Examples of AI “Grandmas” in Action
-
Daisy (UK – O2)
Daisy is an AI voice agent that mimics a friendly but forgetful pensioner. Some fraudsters waste up to 40 minutes before realising Daisy is just code.
Read more about Daisy » -
Apate Bots (Australia – Macquarie University)
These bots take over when a scam is suspected, keeping criminals talking while gathering intelligence.
Read about Apate Bots » -
Lenny (Classic Bot)
A long-standing pre-recorded “elderly gent” who politely rambles until the caller gives up.
Wikipedia Entry »
How AI Filters Genuine Calls from Scams
Modern AI call protection uses multiple layers of filtering:
- Caller Reputation Scoring – Numbers are checked against scam databases and behavioural patterns.
- Real-Time Call Analysis – AI listens for scam scripts, unusual requests, and pressure tactics.
-
Dynamic Call Routing –
- Suspected scams go to the AI “grandma” persona.
- Genuine calls go straight through to the real recipient.
- ASR Jamming – Subtle audio distortions confuse scammer bots but remain clear for real humans. (Research Link »)
Why Wasting a Scammer’s Time Really Works
Every minute spent talking to an AI is a minute scammers can’t use on a real victim. Across thousands of calls, this cuts deep into their profits — and deters them from targeting certain regions altogether.
This method is already proving its worth in the UK and Australia, and businesses like Norango.ai are now adapting it to protect companies as well as individuals.
A Serious Solution with a Sense of Humour
While it’s funny to imagine a scammer stuck listening to an AI describe her grandson’s pigeon-racing club for the third time, the results are no joke. AI voice agents are:
- Reducing scam call success rates.
- Protecting vulnerable people.
- Gathering intelligence for law enforcement.
- Freeing up real people’s time and safety.
Bottom line: The next time a scammer thinks they’ve found an easy mark, they might just find themselves speaking to an AI granny who’s got all the time in the world — and isn’t giving them a penny.
