Zaheer Ebrahim, Solutions Architect at Trend Micro MEA, gives some insight into how this new technology will better support bad actors and how to lookout for what’s real and what’s fake.
It’s predicted that the recent advancements in generative AI will make it harder for people to spot the difference between fake news and misleading information. The World Economic Forum has even declared misinformation powered by generative AI as a top risk facing the world in 2024.
Over the last year, there have been numerous examples of how AI tools are being used to produce fake images and videos of well-known public figures. Images of the Pope Francis in a white puffer coat and former US President Donald Trump being arrested circulated the internet in early 2023. More recently, a sexually explicit deepfake video of Taylor Swift was doing the rounds on X, which caused the social media platform to halt searches for the popstar to prevent the video from spreading. Given how real they can look, AI-generated media used to sway public opinion is cause for concern.
Spot the difference
At first glance these images and videos can seem very real, but upon closer inspection, there are some telltale signs. The first being that AI-generated content has imperfections. This can sometimes be seen in the subject’s hands or in objects that are not fully formed. Often the images seem to have an inconsistent quality and can appear blurry with a glossy effect. In AI-generated videos, people often don’t blink and make unnatural movements.
There are of course programmes being developed and used to detect whether images and videos are real or not, but it will mostly be up to audiences to make up their own minds. So before you forward on that image or video to the next WhatsApp family group or share to social media, consider if it’s AI generated or real. Ask where the video is from and can the source be trusted.
Stay vigilant
While the country gets ready to vote, cybercriminals are considering different ways they can exploit this opportunity. A national election is the kind of major event that cybercriminals will look to gather people’s personal information. And with the help of generative AI, social engineering scams are becoming more sophisticated. In fact, Trend Micro’s latest research has found that AI-generated media will give rise to more polished and persuasive phishing scams.
In the run up to the elections, South Africans are being asked to check that they are registered to vote. The Independent Electoral Commission (IEC) has made it possible for voters to register online via elections.org.za and check where they are registered to vote by simply entering their ID number into the website.
During this time, it’s possible that we will see an increase in SMS and email phishing scams with bad actors posing as the IEC to gather personal information such as ID numbers, home addresses and phone numbers. It’s therefore important to stay safe and be wary of any website, email or SMS that asks users to click on a link and share personal information.
Rather than clicking on a link from a social media site, email or message, type the address into the browser. Clicking on unknown links can send you to a phishing website looking to collect your details.
After typing in the address, double check that the website is secure by looking for a padlock next to the link. You can also see if the website link starts with ‘https’ – the ‘s’ indicates that a site is secure. If you’re still unsure, contact the IEC directly to verify the website address.
In addition to checking the website address and staying away from unsolicited links on email, it’s vital that your device’s security software is up to date. This line of defence will help to scan your device for any malicious activities.
As with any new technology, there is an unknown set of risks. We are learning more about how generative AI can be used by both cyber defenders and attackers. In the meantime, staying one step ahead of these bad actors through constant vigilance can go a long way to being safe and secure online.
