Hype rating
AI usage
Who benefits?
The premise and promise of the effort by Virgin Media O2 (a british telecom service) to “fight” back against phone scammers is appealing, it speaks to anyone who has dealt with unsolicited phone calls, or worse been victims or knows gamily members who have been victims of scams.
Their own article, “O2 unveils Daisy, the AI granny wasting scammers’ time” opens right off with a news story video where we can see a Generative AI characer of an elderly woman is quite compelling in the realism of the video generated character and how she reacts in a befuddled way to the voice insrucions coming in on the line.
This is the appeal:
As part of Virgin Media O2’s ongoing “Swerve the Scammers” campaign, Scambaiter Daisy has been created in response to new research from O2 revealing 7 in 10 (71%) of Brits would like to get their own back against scammers that have tried to trick them or their loved ones. However, not wanting to waste their own time ranked as the top reason why the public wouldn’t bait scammers themselves (53%).
With more than two thirds of Brits (67%) concerned about being the target of fraud and one in five (22%) experiencing a fraud attempt every single week, O2 is fighting back.
The concept is that unsupervised, Daisy can inreract with a spammer in a live conversation, with responses meant to prolong the amount of time a scammer spends wtih Daisy I is very much the premise of the 2015 TEDTalk by James Veitch “This is what happens when you reply to spam email” His counter intuitive strategy was to spend as much time stringing along a spammer, so tha they were thus spending less time abusing other people.
It’s not exactly clear how Daisy works- is Virgin O2 collecting reports of scammers phone numbers, and then calling them back? It’s not clear why a video needed to be generated, except maybe for publicity.
What concerns does this bring up?
There might be a small chance Daisy calls someone who is not actually a scammer. And what is going to stop scammers from using AI for fraud, so the efforts are less scripted by people with not great skills in personas? That might be a stretch,
Additional comments
There’s not much hype here and the details of how Daisy was trained and worked are no described. Again, it’s hard to to see a argument agains the use of AI here to fight fraud.
It does seem, but I cam no sure, that the opening four bullets might an AI gnerated summary of key points.
Also, as a reference, I heard about this story as reference in Mystery AI Hyper Theater’s Episode 47: Hell is Other People’s AI Hype
The show notes link to a post from The Byte called Phone Provider Deploys “State-of-the-Art AI Granny” to Waste Scammers’ Time– this is more or less a summary and in the spirit of following the source upstream, I followed the link to the story referenced here.