More scam calls relying on Kiwi Fomo and deepfakes getting complex, cyber agency says

The misuse of artificial intelligence (AI) is expected to cause more sophisticated and dangerous scams, a cyber security agency says.

Cert NZ’s incident response team leader Tom Roberts said AI’s ability to replicate voices, natural language and even realistic video would lead to more online fraud.

TVNZ presenters target="_blank">Wendy Petrie and Simon Dallow have had their images misused in AI-related scams recently.

“If you find your image or profile is being used in a scam, immediately report to the platform it’s on,” Roberts said.

The fake Petrie ad had 1News video interspersed with file footage of a Ferrari and money but with a strange and jumbled accent used to voice the scripted gambling ad text.

“For those who see these ads online, the best defence is common sense,” Roberts said.

A scourge of pornographic deepfake images generated by artificial intelligence has affected public figures including singer Taylor Swift. Photo / AP

“Think about whether a celebrity would actually be advertising the product it claims they are.”

He said AI scams still relied on social triggers such as urgency and the fear of missing out (Fomo).

“Be sceptical of any offers and also check the official page of the celebrity to see if they have shared the ad as well.”

Meanwhile, Cert NZ said it had detected more scam calls such as a fake Visa racket recently operating from Australian phone numbers.

“In general, the scammers are trying to get as much information from the target as possible, and then they will use that information in other scams,” Roberts said.

“Hang up. If you’re concerned that the call may be legitimate, it’s much safer to hang up and call the organisation back on their official phone number, be it your bank or a government agency.”

He added: “If you realise too late that it’s a scam and have given over details, report it to Cert NZ and the relevant organisation, especially if it involves your bank details.”

Rapidly advancing technology misused for scams or revenge porn can be prosecuted under existing laws, Justice Minister Paul Goldsmith said.

Goldsmith said AI and deepfakes presented some challenges.

“Technology has advanced in such a way that a person does not have to take a photo or video for one to be used against them,” Goldsmith said.

Justice Minister Paul Goldsmith says existing laws could be used to prosecute people responsible for sexualised deepfakes. Photo / Mark Mitchell

“Deepfakes are particularly harmful when they are of a sexual nature.”

Globally, victims of such deepfakes included singer Taylor Swift.

Goldsmith said Crimes Act offences already related to intimate visual recordings, including prohibitions on making them, possessing them in some circumstances, and publishing, importing, exporting, or selling them.

“The Harmful Digital Communications Act 2015 also provides an offence for posting an intimate visual recording without consent,” Goldsmith added.

“I’m advised that in most circumstances this act will determine posting deepfakes of a sexual nature to be an offence.”