Telecoms companies now using scambaiting AI systems modelled on Lenny

A new breed of scambaiting AI systems are now being officially used by telecoms companies to protect their customers from fraud. The systems use more sophisticated versions of Lenny, a delightful simulation of a rambling old man keen to engage in lengthy conversation …

Scam calls are now a huge problem

A recent survey suggested that scam calls now cost Americans a total of more than $25B per year. The average loss is $452, but some victims have lost their entire live savings – or even gone into debt to make what they thought were legitimate payments.

Romance scams are a common example. This is where the scammer romances the victim on a dating app, gradually gains their confidence, then claims to need a short-term loan to fund a plane ticket to meet them, or to cope with an emergency.

Another common one is a called claiming to be from the IRS chasing an overdue payment which the victim must pay immediately to avoid being arrested.

Fake purchases are also common, where an email is sent with a receipt for an expensive purchase, with a number to call to arrange a refund. The aim of these is to obtain bank details that are then used to take money from the account.

Scambaiters fight back using Lenny and more

Unsung heroes known as scambaiters have been fighting back for some years now. At the simplest level, they aim to keep scammers on the phone for as long as possible, in order to reduce the amount of time they have available to make other calls.

In the most sophisticated of cases, scambaiters will obtain details of bank accounts used to launder the money, in order to get these accounts shut down.

One tool used by scambaiters is known as Lenny – a series of voice messages which automatically play whenever the scammer stops speaking. Lenny rambles on at great length, and has proven remarkably effective at keeping scammers on the line for extended periods. If you’ve never listened to one of these calls, I highly recommend them! You can listen to one example below:

Telecoms companies now using AI chatbots

The Guardian reports that telecoms companies are now starting to use hundreds of AI chatbots – essentially more sophisticated versions of Lenny. One of them is named Ibrahim.

Ibrahim, a cooperative and polite man with an Egyptian accent, picks up. “Frankly, I am not too sure I can recall buying anything recently,” he tells the hopeful con artist. “Maybe one of the kids did,” Ibrahim goes on, “but that’s not your fault, is it?”

[He’s just one] of the conversational artificial intelligence bots created by Prof Dali Kaafar and his team. Through his research at Macquarie University, Kaafar founded Apate – named for the Greek goddess of deception.

Apate is designed to do more than just waste the time of scammers – the AI system also seeks to learn about the scam techniques used in order to help warn people against them, and provide intelligence for law enforcement.

If a telecommunications company detects a scammer and diverts it to a system like Apate, the bots will work to keep the scammers busy. They test different strategies, learning what works to make sure scammers stay on the line for longer. Through success and failure, the machines fine-tune their patter.

As they do this, they extract intelligence and detect new scams, collecting information on how long the call lasts, when the scammers are most likely to call, what information they are after, and what tactics they are using.

While scambaiting can be fun, ordinary consumers are advised to simply hang up.

Photo by Alex Knight on Unsplash