AI has a lot of potential uses, many of them bad and many good. Today’s case is one of the good ones. Or rather, one of the brilliant ones, as creating an avatar of a grandmother managed by AI to drive phone scammers crazy is brilliant. Let’s tell the story.
Subscribe to the Softonic newsletter and get the latest in tech, gaming, entertainment and deals right in your inbox.
Subscribe (it's FREE) ►Daisy, or dAIsy, is an AI created by O2, the largest mobile operator in the United Kingdom, to trick scammers into believing they are talking to a real grandmother who likes to go off on tangents. When the AI provides the requested bank details, it reads out loud fake numbers and names.
And the software is designed to keep people online as long as possible. This not only means less time for scammers to target real people, but O2 also uses the conversations to learn the favorite tricks and techniques used in these scams.
As you can hear in the video, the scammers are not happy about being tricked: they get increasingly angry and start swearing. The bot is so convincing that it has managed to keep some people on the phone for 40 minutes straight.
If you have seen any of the several YouTube channels that scam scammers, sometimes using a voice changer to sound like an elderly woman, you will know what to expect. Daisy has been trained with the help of one of the most popular scammers on the platform, Jim Browning.
How Daisy works, the Grandma who trolls scammers
Daisy listens to the caller and transcribes their voice to text. The responses are generated through a custom LLM with a personality layer (of a grandmother in this case) and are then returned through a custom AI text-to-speech model to generate a voice response. All in real-time.
O2 customers do not have access to Daisy to launch their own revenge campaign against scammers. Instead, the AI tool has been added to a list of “easy target” numbers used by scammers. Daisy is capable of interacting with callers 24/7 without human operator intervention.
Daisy was created in response to an O2 study which found that 71% of Britons would like to get back at scammers who have deceived them or their loved ones, but most said they wouldn’t engage in scamming because they didn’t want to waste time.
Although the work done by AI should be applauded, its ability to converse with someone so convincingly is unsettling. Ironically, scammers also use similar technology to trick people into believing they are talking to their relatives. In this war, we only wish for the scammers to lose.
