TECH-savvy criminals have been latching onto the advancements of artificial intelligence and attempting to con millions of people online with increasing regularity.
While the likes of ChatGPT have shown promising signs of adding efficiency and information into our lives like never before, the doors have also been opened for scammers using AI to wreak havoc on unsuspecting victims.
3
3
Highly skilled hackers have created “bad bots” like WormGPT, WolfGPT, and FraudGPT to roll out phishing attacks and malware.
Thanks to the ability to analyze vast amounts of data, highly personalized and convincing scams attempt to steal bank details or push people into clicking on links, opening the door for large-scale attacks.
They can also identify vulnerable websites for hackers to target and generate code to create new forms of malware with the potential to evade traditional detection.
Social media scams, voice cloning, and the rise of highly sophisticated deepfakes have the potential to dupe unsuspecting members of the public, with elderly tech users most at risk.
Deepfakes – computer-generated copycats designed to mimic the look and sound of well-known people or celebrities – have been identified by experts as a serious problem.
The victim may instantly trust the online presence because it’s someone they recognize, and therefore have less of a problem transferring money or handing over valuable pieces of personal information.
Former Presidents Barack Obama and Donald Trump, as well as current POTUS Joe Biden have all been impersonated, while Hollywood stars like Leonardo DiCaprio and Tom Cruise have also had their images represented with the intent of tricking people.
However, there is hope in the battle to foil the dark arts.
Most read in News Tech
Marijus Briedis, chief technology officer at cybersecurity company NordVPN, says the hackers are still prone to errors that could give their game away.
“Look out for sudden head movements that can reveal blurry edges to the figure on film, or unusual changes to the lighting,” he told The U.S. Sun.
“The way an AI deepfake is grafted to the image of the real person can be startlingly smooth, making the final result seem shiny, or as if a social media filter had been applied to it. Mouth movements may also appear unnatural to the naked eye.”
Unfortunately, deepfakes are just part of the growing problem in the battle for online security.
Fake LinkedIn and social media profiles are everywhere.
A common approach is someone being contacted by a fake profile that pretends you have people in common.
A bot, who is pretending to be a real person, will try to get in touch and attempt to convince you to sign up for an investment scam.
“On professional networking sites such as LinkedIn, people are often more likely to accept requests from people they don’t know, especially if they are in the same industry or a position to help them in their career, such as a recruiter,” said Breidis.
Recent research from NordLayer showed that 47 percent of LinkedIn users in the United States have received a request to connect with a suspicious link and 29 percent received an invitation to take part in a fake survey.
LinkedIn, meanwhile, identified over 21million fake accounts in the second half of 2022, suggesting one in 40 users were bots.
The profiles often offer jobs that don’t exist.
“A fraudulent LinkedIn profile will usually make the first move to contact you, so be very cautious about clicking on links or when being told about job offers,” added Briedis.
“The scammer’s profile can usually give away hints that they might not be who they say they are. Real profiles link to the person’s company website or their professional social media profiles, but be careful clicking any links that have been shortened, as it could be another way of infecting your computer with malware.
“Look for inconsistencies in profile posts such as spelling errors, and be wary of profiles with a whirlwind of activity in a very short time period, or if they have very few connections.
“Thankfully, LinkedIn has stepped in to address this type of scam, announcing recently that it had developed a new AI image detector to catch fake profiles, with a 99 percent success rate.”
AI has also been created to produce highly personal voice cloning scams where the sounds of someone talking can be cloned before being released to unsuspecting people.
A child’s voice, for example, could be used to trick a parent into parting with large sums of money.
Scam calls do the groundwork, asking people if they can hear them before recording their voice and then letting the AI do the rest of the dirty work.
“Voice cloning scams are a sick imitation game capable of cheating people out of thousands of dollars,” said Briedis of the rising threat of identity fraud.
The best way to weed out any wrongdoing is, according to Briedis, to ask personal questions that the AI will have no chance of knowing, as well as to review privacy settings in apps that use voice data and opt out of data voice collection.
Other important tips include avoiding posting personal information online, always considering how necessary it is to share details, and always ensuring all security updates are used whenever available.
3
