DEEPFAKES: THE DECEPTIVE ILLUSION

We live in a world where we are constantly bombarded with imagery and sounds in our daily lives. Since most of us spend a good amount of time on the internet, it is sometimes hard to distinguish what we see or hear is real or fake. With the advancement of Artificial Intelligence, using DeepFakes, it has become even more difficult to discern what is real or fake. So what are DeepFakes?

WHAT ARE DEEPFAKES?

DeepFakes use a form of artificial intelligence called deep learning to make images of fake events. People will use Artificial Intelligence Programs to create visual pics or audio in order to spread disinformation or used in shilling. DeepFakes was started in 2017 when a Reddit user of the same name posted doctored porn clips on the site.

DeepFakes transform existing source content, such as replacing one person for another. DeepFakes can be pics, video, audio recordings, which can look and sound real, but have been manipulated via Artificial Intelligence Programs in order to make people appear to do or say things they never did or said. Examples of DeepFakes are: Face Swapping, Lip Sync, Puppet Master( replacing one person’s body image with another body image).

WHO MAKES DEEPFAKES:

Basically anyone from academic and industrial researchers to amateur enthusiasts, visual effects studios and porn producers can make Deep Fakes. Governments might also use the AI technology to discredit and disrupt extremist groups and to target individuals directly.

DEEPFAKE IN AUDIO:

Audio can be DeepFaked too, to create “voice skins” or ”voice clones” of public figures. This can create problems when the public figures are looked up to. The Audios are created to discredit the public figures and tarnish their reputations and create doubt with those who support them. DeepFake audio is used often for scammers who are trying to con people via the telephone and/or online audio chats.

DEEPFAKE IN VIDEOS:

DeepFake Videos primarily use face swapping. First, AI programs will run thousands of face shots of the two people through an AI algorithm called an encoder. The encoder finds and learns similarities between the two faces, and reduces them to their shared common features, compressing the images in the process. A second AI algorithm called a decoder is then taught to recover the faces from the compressed images. People who are engaged with DeepFake Videos will use them to discredit, con, and persuade others to see someone in a negative light, push a narrative or an agenda. They see many people are not paying attention and question what they see. This gives them an advantage. To check to see if a video is DeepFake you can use various web applications such as DeepWare.

DEEPFAKE IN PORNOGRAPHY:

A majority of DeepFakes are done in the Pornography Industry. In 2019, there were 15,000 DeepFake videos online and every six month since then, the numbers have doubled, which is quite staggering. 96% of DeepFakes are pornographic and 99% of them are face swapped faces from female celebrities on to porn stars. This is just in the adult porn industry. The Child Porn industry it is even worse. It is used to exploit children over and over again without recourse. Law Enforcement will sometimes use DeepFakes to bait potential predators of children in order to make arrests. However, in some states, DeepFakes are illegal, such as Virginia, Texas, and California.

SPOTTING DEEPFAKES:

Poor-quality DeepFakes are easiest to spot. The lip synching is bad, the skin tone patchy, there is flickering around the edges of transposed faces. Fine details, such as hair, are rather hard for DeepFakes to render well, especially where strands are visible on the fringe. Badly rendered jewelry and teeth are also a dead giveaway. Plus lighting effects, such as inconsistent illumination and reflections on the iris, can make it easy to spot DeepFakes. Also, unusual skin tones, stains, strange lighting, and oddly positioned shadows may indicate seeing a DeepFake. If watching a suspicious video, there will be discrepancies in the person. Best to compare the DeepFake to an original reference.

EXAMPLES OF DEEPFAKES:

1. Donald Trump, who admitted to boasting about grabbing women “by the pussy,” in a recorded conversation, was not even real. It was a DeepFake created to discredit him when he was running in the 2016 Presidential Election.

2. In a BBC interview with Emily Maitlis, Prince Andrew cast doubt on the authenticity of a photo taken with Virginia Giuffre, a shot her attorney insists is genuine and unaltered. It turned out to be a DeepFake.

3. Jim Carrey’s brief and unexpected appearance in The Shining is anything but hilarious. It was a well rendered fan-made DeepFake. Carrey’s facial expressions were quite terrifying, impersonating Jack Nicholson in the film’s typewriter scene.

4. Recently posted pic of Donald Trump kneeling and praying, which hit social media platforms in March 2023, became a hit is also a DeepFake.

   

5. In 2018, Buzz Feed Video made a DeepFake of former President Barack Obama. The DeepFake accurately and perfectly mimicked Obama’s voice and gestures to a point where one could not tell the video was DeepFake. Towards the end of the video, it was revealed that actor, Jordan Peele, was the one impersonating the former president.

6. In March 2023, a high profile group, on the social platform Truth Social and Telegram, posted a pic with President Trump at Mar A Largo. The pic looks like President Trump; however after further investigation, the photo is another DeepFake. It was done by AI. These accounts made another DeepFake of a vehicle with the the groups logo on it.

DEEPFAKES AND EMPLOYEE FRAUD:

DeepFakes are not all about fun and games. Numerous safety risks come with DeepFake technology. It can be potentially used to commit fraudulent acts. DeepFake threats are able to target the workplace. Online impostors are able to use real-time DeepFake videos to extract job candidate information during remote interviews. This allows them to gather personally identifiable information (PII) through posting fake job openings and screenings. Imposters desire money, access to company systems, espionage, and unearned career advancements. These factors explain why the usual jobs impostors apply for are IT and computer programming, database, and software-related job functions.

PROTECTING ONESELF FROM DEEPFAKES:

With the advancement of Artificial Intelligence, and the increase of DeepFakes, there are ways to protect oneself. Several ways to ensure safety are:

1. Make sure family knows about how DeepFakes work and the challenges it can pose.

2. Educate yourself and others about how to spot DeepFakes.

3. Keep up on media literate and use good quality news sources.

4. Trust but Verify. Keep a skeptical attitude. This will help you avoid many traps.

DeepFakes can be be deployed by hackers attempting to break into home and business networks. The best way to minimize the risks are:

1. Make sure you protect your data against ransomware. It will give you the ability to restore damaged data.

2. Make sure they are strong password. Use different ones for different accounts. This means if one network or service is been broken it will not affect the others. If one is compromised, the others will not be.

3. Find a good security package to protect your home network, laptop and smartphone against cyber threats. Make sure it as VPN and anti-virus software.