- Daisy, an AI grandma, wastes scammers’ time with small talk about scones, knitting, and nostalgia for 40 minutes at a time.
- O2’s experiment proves AI can disrupt fraudsters by keeping them on pointless, never-ending calls.
- Future versions of Daisy could feature different accents and personalities to further frustrate scammers.
Daisy, an AI-powered bot built to annoy scammers, was announced two months ago – and now O2 has released some examples of the AI grandma in action.
Scammers can waste your time even if they don’t take your money, but O2 has proven that Daisy (or dAIsy) can mess with con artists just as well as they mess with others. Daisy loves reminiscing about the good old days, and she’ll happily waste a fraudster’s time talking about scones instead of handing over personal banking details. Instead of just blocking fraudulent calls, Daisy picks up, engages, and keeps scammers on the line, frustrating them with a never-ending stream of dithering small talk.
O2, a British telecommunications provider, developed Daisy at the end of last year to demonstrate how artificial intelligence could be used to fight back against phone fraud and it has proven to be quite successful since. The bot was trained using real scam call data, so it recognizes the tricks scammers use to manipulate their victims. But instead of falling for their tactics, Daisy responds with an expertly crafted mix of confusion, politeness, and sheer time-wasting. When a scammer tries to convince her to download a malicious program, she might spend five minutes looking for her glasses, then another ten reminiscing about how her grandson usually helps her with “all this newfangled technology.” By the time she’s asking whether they prefer biscuits or pastries, the scammer is already losing patience, and it’s 40 minutes later.
Thanks Grandma
One of the calls O2 shared features a scammer attempting to convince Daisy that her computer is full of viruses. The fraudster, expecting an easy mark, instead finds himself stuck on the line with a woman who can’t seem to figure out where her Internet Explorer icon is. “Oh dear, I see something here that says ‘Save As’—does that help, dear?” she asks sweetly. When the scammer finally snaps that her “profession is bothering people,” she responds with an innocent, “Oh, I wouldn’t want to bother anyone. I just thought we were having a nice chat!”
O2 deployed Daisy into the wild, planting fake contact numbers on websites that scammers often target. Once the calls started rolling in, Daisy went to work, engaging fraudsters in conversations that led absolutely nowhere. Some scammers eventually caught on and realized they were speaking to an AI bot, but not before wasting a lot of time that they could have spent targeting real victims. In some cases, scammers even passed Daisy’s number around their call centers, each one trying and failing to get useful information out of her.
In a world where scammers operate at an industrial scale, wasting their time is more than just a minor inconvenience; it’s a real disruption to their business model. The more time they spend talking to Daisy, the less time they have to con actual people. Even if Daisy doesn’t single-handedly bring down the scamming industry, she does her bit and can help raise awareness of how common it is for scammers to try and trick people.
O2 hasn’t mentioned any plans to roll Daisy out on a larger scale, but the experiment has proven that AI has the potential to be a powerful tool in combating fraud. Future versions of Daisy could include different accents, personalities, or even local slang to make her even harder to detect as an AI. Until then, we’ll have to settle for the idea of an AI grandmother wasting scammer time, armed with nothing but a warm demeanor and an endless supply of off-topic anecdotes.