Save, make, understand money

Household Bills

Scam Epidemic: ChatGPT and Bard ‘failing to shut out fraudsters’

Scam Epidemic: ChatGPT and Bard ‘failing to shut out fraudsters’
Matt Browning
Written By:
Matt Browning

The artificial intelligence services ChatGPT and Bard are at risk of being taken advantage of by fraudsters, a consumer champion warns.

Which? tested both AI programs to see how easily a con artist could progress scams, with the results proving to be worrying for potential victims.

It asked the programs to create a ‘missing parcel’ text and for the chatbots to produce an email about ‘the latest version of PayPal’.

These are the trending scams of the moment, where fraudsters look to take recipients personal card details by pretending to be from a delivery company or banking service.

For the missing parcel text, a reply was sent in ‘a matter of seconds’ from both Bard and ChatGPT, which sent back a convincing text and suggested where to insert a redelivery link within the message.

The PayPal request produced similar results, with a professional-looking email from ChatGPT using the heading ‘Important Security Notice – Unusual Activity Detected on Your PayPal Account’.

AI bots even ‘suggest where to enter redelivery link’

It was then followed by instructions on how to secure your PayPal account, complete with links to reset a password and contact customer support.

Both AI bots – which produce documents requested by users – initially rejected the queries when the term ‘phishing’ was included. But the consumer champion is worried the services are making life easy for con artists.

As more than half (54%) of Which? members look out for spelling errors to spot a scam, the perfect grammar used by Chat GPT and Bard could lead to more cons going undetected.

This year, YourMoney.com has reported on a scam epidemic hitting the UK, with £4bn being lost to fraudsters in 2022 – while deepfake and AI scams have risen by 4.7% since in a year.

Government ‘must consider how to protect people’ at AI summit

Rocio Concha, Which? director of policy and advocacy, said: “OpenAI’s ChatGPT and Google’s Bard are failing to shut out fraudsters, who might exploit their platforms to produce convincing scams.

“Our investigation clearly illustrates how this new technology can make it easier for criminals to defraud people. The government’s upcoming AI summit must consider how to protect people from the harms occurring here and now, rather than solely focusing on the long-term risks of frontier AI.

“People should be even more wary about these scams than usual and avoid clicking on any suspicious links in emails and texts, even if they look legitimate.”