Of all the AI chatbot ideas we have seen since ChatGPT went viral nearly two years ago, O2’s Daisy (stylized as “dAIsy”) is easily among the best. It’s an incredibly brilliant idea from a British operator who wants to prevent scammers from reaching. it does this by having Daisy and its lovable grandma personality waste scammers’ time by keeping them on the line until they can’t take anymore.
That’s a great solution to a problem that’s not going away anytime soon. Scammers will call as many people as they can with the hope of scamming them out of money one way or another. They use all kinds of pretenses to convince people to hand over their personal information.
Some of these scams might not necessarily go for your bank account. Instead, they might try to take over certain account logins. Some scammers might even use AI to give you the impression you’re talking to a real person from a customer support department.
The elderly are the easiest prey, at least in theory. Not all of them know how to deal with scammers, and some of them might be easily tricked. That’s why Daisy is such a brilliant idea.
By giving this chatbot the voice and tonality of a nice elderly lady who has all the time in the world to chat, O2 practically guarantees the scammer will stay on the line. They’ll be convinced they can get results, no matter how long it takes.
As you’ll see in the video below, Daisy tricks scammers into spending time with her on the phone. The clip contains excerpts from real chats with scammers that lasted 40 minutes or more. The visual representation of a grandma only benefits this promo clip. In reality, Daisy doesn’t have a face, and scammers will never see her.
The longer the scammers talk to Daisy, the better for real people. The scammers would not be able to target others if they were busy trying to convince Daisy to give up her credit card information or whatever they might be trying to obtain.
It gets even better than that. If and when the scammer finally gets a credit number from Daisy, it won’t be real. But they’ll spend even more time trying to get something out of it.
Projects like Daisy are possible thanks to the massive advancements in genAI software in recent years. Chatbots can talk about anything, and thanks to voice features like Advanced Voice Mode (ChatGPT) and Gemini Live (Gemini), they can hold voice conversations that sound exactly like a talk between humans.
O2 doesn’t say which genAI models it’s using, but the granny chatbot is “rained using cutting-edge technology and real scambaiter content.” As a result, Daisy is “indistinguishable from a real person,” which works out great for the purpose of this AI use:
Daisy combines various AI models which work together to first listen to the caller and transcribe their voice into text. Appropriate responses are then generated through a custom large language model complete with a character ‘personality’ layer, and then fed back through a custom AI text-to-speech model to generate a voice answer. This takes place in real-time, allowing the tool to hold a human-like conversation with a caller.
How do scammers reach Daisy? O2 added the AI’s number to “easy target” lists that scammers would have access to. O2 could always keep things interesting by changing the number or adding multiple lines. That’s just speculation, however.
As for that real scambaiter content, O2 says it worked with Jim Browning, one of YouTube’s best-known scambaiters. The carrier says it created Daisy after finding that more people would want payback against phone scammers. O2 research showed that 7 in 10 Brits would want to fight back. At the same time, 53% of them do not want to waste their time keeping scammers busy.
So far, Daisy is working as intended, but the AI hasn’t been up for too long:
After several weeks of taking calls in the run up to International Fraud Awareness Week (November 17-23), the AI Scambaiter has told frustrated scammers meandering stories of her family, talked at length about her passion for knitting and provided exasperated callers with false personal information including made-up bank details.
By tricking the criminals into thinking they were defrauding a real person and playing on scammers biases about older people, Daisy has prevented them from targeting real victims and, most importantly, has exposed the common tactics used so customers can better protect themselves
Daisy will go viral, and scammers will become aware of its existence. However, they won’t be able to determine when they’re talking to AI or a real person as they start a call.
Some of the more resourceful scammers might employ AI chatbots of their own to hold these conversations. But even then, they’ll waste time when dealing with Daisy.
Separately, O2 is encouraging Brits to report scam numbers at 7726. After that, O2 can block some attacks. The carrier blocked 89 million texts last year, thanks in part to the service.
