Deepfakes: Is the future fake?
Date: Jan 16, 2022
What is a deepfake?
Deepfakes are an ever-emerging technology. The use of artificial intelligence to alter and recreate digital content to make it appear as if something that didn’t happen did. Deepfake software use large amounts of data to introduce existing audio and visual content into other media, producing an altered but seemingly real video or image.
Deepfake in Cybersecurity:
Starting out deepfakes were first used for fun. Originally used in 2017 on Reddit, by a user named “deepfakes”, who posted videos of celebrities faces swapped onto the bodies of actors in pornographic videos. As time goes on and deepfakes get more and more sophisticated, deepfakes begin posing a real threat in both the real and cybersecurity world.
In 2020 The Verge wrote about a deepfake audio scam. Where a hacker made a deepfake recording of a CEO asking an employee to urgently assist him in finalizing a business deal via voicemail. Fortunately the deepfake wasn’t too great and the employee realizing something was wrong, forwarded it to the legal dept.
But good audio deepfakes can be successful like in 2019 when a CEO of a U.K.-based energy firm thought he was speaking to his boss and was convinced to wire $243,000 to hackers.
Lately deepfakes have been getting more and more accurate. An imposter, claiming to be Andre Conje, the creator of yearn.finance scammed the Blackswan token team using a deepfake resulting in a loss of $30,000 worth of tokens. Using widely available videos of him, the imposters created a deepfake zoom video call impersonating Andre.
Countermeasures:
Of course, the security world is not sitting idly by as hackers think of new ways to compromise our systems. Various AI intelligence tools have come out overtime to detect deepfakes, such as Deepware. Deepware allows you to scan and detect deepfakes through either uploading or inserting a link to videos. The problem with AI tools is the constant advancement of the deepfakes. These tools have to be continuously updated with the latest deepfake detection technology to be able to accurately identify deepfakes.
Like phishing, to really prevent deepfake threats, businesses need to create awareness and put in place prevention measures. Employees should be trained to look out for robotic sounding voicemails, or suspicious looking videos. Furthermore, all important correspondence must be authenticated in more than one way.Deepfakes are a growing threat to the security world. Businesses must be one step ahead, and constantly be on the lookout to where it may affect them.