Daisy Harris likes birds: there is one outside her window. Her cat is called Fluffy and she would love to tell you about her love of knitting. He likes tea and cookies. And the truth is that he doesn’t understand how the internet works.
But maybe you can explain it to him just one more time?
More or less this would be a conversation with Daisy Harris, a grandmother generated with the help of artificial intelligence presented this month by the large telephone company british O2, as part of its efforts to stop phone scammers. He has no badge, no warrant, no way to actually stop them. But when an unlucky scammer dials her number, Daisy has the power to waste an unlimited amount of his time.
There are many real people who like to set up scammers: they identify potential thieves and give them a soup of their own chocolate, initiate endless conversations so they can’t call other potential victims. Unlike them, Daisy has no impediments, such as the need to sleep.
“These people can’t talk to thousands of scammers,” said Morten Legarth, who helped develop Daisy with VCCP, a London advertising agency. “But there is this idea that AI can.”
Phone scams have reached mind-boggling levels: Last year, the number of fraudulent calls soared into the tens of millions, according to phone security company Hiya. An anti-scam consortium claims more than $1 trillion was stolen, most often when victims unsuspectingly provided banking details, passwords or other personal information.
The Internet has only made these scams easier and, although scammers do not discriminate, older adults are easy prey: In a British study, 40 percent of those over 75 reported receiving scam calls at least once a month, if not daily.
Daisy, with her bewilderment at technology and her desire to please, presents herself, at least at first, as the perfect target. Its developers explained that they were based on expectations, often taking inspiration from their own grandmothers.
“I was very inspired by my grandmother. He always talked about the birds in his garden,” recalled Ben Hopkins, who also worked on the VCCP project. Instead of using a voice actor to train Daisy, the team opted to use one of their teammates’ grandmother, who came over for tea and recorded hours of dialogue.
A prolific Northern Ireland-based scammer who posts YouTube videos under the name Jim Browning worked with O2 and VCCP on the development of Daisy, providing it with techniques to keep scammers on the phone. Among them: going off on tangents with topics like hobbies and family, and feigning technological ineptitude.
In one case, three phone scammers joined in on a call that lasted nearly an hour, trying to get Daisy to type “www.” in a web browser.
Still, given the sheer volume of scam calls, Daisy’s efforts are far from a roadblock and are just a small obstacle. O2 customers cannot forward suspicious calls to Daisy; rather, she answers a few different phone numbers from the immense haystack of digits that scammers can dial.
“Although, in practice, it interrupts operations, it does not stop fraud in a broad sense”said Elisabeth Carter, associate professor of criminology and forensic linguist at Kingston University in London.
Carter advised against trying to follow Daisy’s example of fighting scammers. As satisfying as it may seem, “the best thing you can do if you receive a call from a scammer is not to get involved, hang up and report it,” he advised.
© The New York Times 2024.
