A deepfake is a seemingly realistic video, picture, audio, or text that’s made using artificial intelligence (AI). As deepfake technology has become more accessible, cybercriminals have begun harnessing it at an alarming rate, including to perpetrate identity theft. Combat it by using strong security settings for your social media accounts and any other websites containing your personal details.
In January 2024, a new comedy special by George Carlin hit YouTube, with the famous comedian turning his signature wit to topics like artificial intelligence (AI) and streaming services.
Carlin, however, died in 2008.
The show was a deepfake and Carlin’s estate sued the project’s creators for copyright infringement.
Some deepfakes are generated for harmless fun, but they’re more often created solely to exploit.
In the identity theft landscape, deepfakes present a rapidly growing threat. At Allstate Identity Protection, we’re here to make sure you’re up to date on the ways AI-generated technology can be used to impersonate you or to draw you into a scam by impersonating someone else.
What is a deepfake?
Deepfakes are media files created using computer systems that recognize patterns in data — whether images, videos, audio, or text.
To make a deepfake photo or video, hundreds or thousands of images are fed into a network to “train” it. Using the patterns it identifies, the software constructs a replica of a designated subject, such as a person’s face.
The more input the system gets, the more convincing the output can be. In the case of the Carlin special, for example, the project’s creators reportedly fed their network 50 years’ worth of Carlin’s comedy routines so that it could best mimic his comedic style.
This form of machine-learning is called “deep learning,” thus the term “deepfakes.”
Various technologies used to create deepfakes can place one face onto another body, make it look and sound like a person said something they didn’t, or make someone appear to move in ways they didn’t. Software can even clone a person’s voice to leave a voicemail or have a live conversation.
Likewise, it can also imitate an individual’s writing style for emails and text messages.
Deepfakes are increasingly used for identity theft
As the technology involved in creating deepfakes has become more accessible to the masses — and more sophisticated — cybercriminals are ramping up their efforts to weaponize it.
According to a May 2024 Javelin report, 29 percent of businesses worldwide were impacted by video deepfakes in 2022, and 37 percent of them by auditory ones. The next year, deepfake attacks spiked by 3,000 percent.
These corporate attacks can put anyone at a heightened risk for identity theft, as some are designed to gain access to a company’s sensitive (and highly valuable) customer data.
But do you actually need to be concerned about a fraudster making a deepfake of you? Unfortunately, the answer is a louder and louder “yes.”
“The truth is that these technologies can scrub data (such as video, photographs, and voice recordings) of millions of people from websites, like social media platforms,” states the National Cybersecurity Alliance (NCA).
Some cybercriminals can use this data to generate deepfakes that can bypass authentication systems like those often relied on by financial institutions.
For example, your bank may employ facial or voice recognition to verify your identity before granting you digital access to your account. A bad actor with sophisticated software and a good deepfake image could get into your account and authorize fraudulent transactions.
On the flip side, identity thieves are turning to deepfakes to lure victims into schemes such as romance scams. They may use programs to change their voice and facial features, creating a persona that can gain your trust through video calls.
Meanwhile, deepfake voicemails and text messages are adding credibility to phishing attempts.
Keeping your identity safe from deepfakes
As of now, there is no federal law that explicitly bans deepfakes. However, several states have passed laws targeting those who create and share explicit deepfake content.
The good news? There are several things you can do to make it hard for a cybercriminal to replicate you in a deepfake.
Keep images of yourself private. If you share photos and videos of yourself on social media, adjust your account settings so that only people you trust can see what you post. Be cautious about who friends or follows you.
Using watermarks. Placing a digital watermark on any photos and videos deters cybercriminals by making the images easier to trace. This can be done using graphic design software such as Canva.
Enabling strong privacy settings on websites you use. All kinds of sites contain personal details about you, including those for photo storage, e-commerce, email, food delivery services, and online dating. Deepfake creators can target you based on your interests. To change your site settings on Google, start here.
Learn the signs of deepfake media
Staying savvy about AI-generated content and harnessing a healthy sense of skepticism can also help you spot a scam.
A video might be a deepfake if you notice:
Blurriness in just one part of the image or video, such as the face
Changes in skin tone
Double chins or eyebrows, or a double edge around the face
Inconsistencies in the video’s quality
Square shapes around the mouth, eyes, and neck
Unnatural movements of the face or body, including weird blinking, or a lack of blinking
Background or lighting changes or inconsistencies
Audio or text might be a deepfake if you experience:
Received from an unfamiliar name, phone number, or email address
Choppiness in the sentence structure
Strange phrasing or context; if the contact claims to be someone you know, consider whether their wording and topic of conversation feels authentic
Misspellings in text
As identity thieves continue to sharpen their deepfake tools, it’s more important than ever to follow best practices for digital safety. Make sure you’re using strong passwords and set up two-factor authentication (2FA) whenever possible
If you’re an Allstate Identity Protection member and have concerns about deepfakes, give us a call using the number on your account dashboard. Our identity specialists are available to answer your questions 24/7, and if there’s any sign of identity theft, we’ll jump into action to investigate.
