There’s an old adage that “seeing is believing.” However, in the 21st century, emerging technologies continue to blur the boundaries of reality in the digital realm, meaning not every photo or video clip you see today can be assumed true.
Specifically, deepfakes are becoming more common, cluttering the Internet with a deceiving mix of real and artificial content. That’s why 77% of Americans are in favor of restrictions on altered images and videos, according to a survey conducted by Pew Research.
As you continue reading below, we’ll take a closer look at the rising threat of deepfake photos and videos, where you’re most likely to come across deepfakes, and some tips for spotting them.
What’s a Deepfake?
Deepfakes are a type of photo or video that have been digitally altered or generated using artificial intelligence (AI) technology.
It’s a relatively new capability that has been developing over the past decade, using neural networks and facial recognition to take existing source materials (like actual video or photos of someone) to create original, fabricated content.
The technology behind deepfakes is quite complex and sophisticated, though it’s now relatively accessible and easy for people with minimal technological skills to create.
The Deepfake Risk of Misinformation
Deepfakes are meant to be highly realistic and accurately replicate human facial expressions, speech, and movements, which can make it difficult to distinguish between a photo or image of something that actually occurred and one that was artificially produced.
Thus, while not inherently malicious technology, deepfake images and videos can be extremely misleading without the proper context.
Bad actors can create deepfakes to spread misinformation while appearing to come from a trusted authority. Deepfakes can also help support a hoax by duping viewers into thinking that a person has said or done something they haven’t.
Today, deepfakes remain largely legal, and legislation has not yet emerged to fully regulate their use and distribution unless they involve defamation or hate speech. There’s still no clear roadmap or legal recourse for fighting misinformation created by deepfakes, though this will likely change over the coming years with more advanced deepfake detection tools and future legislation.
Common Uses of Deepfakes
Nearly anyone online today can create a deepfake. They might range in level of believability, though all Internet users should be aware of where they might encounter deepfakes to avoid falling for these often hyperrealistic spoofs.
Here are some of the most common uses of deepfakes online today (both innocent and malicious):
Social media
Deepfake content commonly appears on social media platforms. These sites are used by a significant portion of the world’s population, so fraudsters can use this reach to influence and sway large audiences with deepfake content.
According to some estimates, around 500,000 video and audio deepfakes were shared on social media in 2023. By 2025, this number is expected to be closer to 8 million.
The nature of the deepfake content found on social media can vary greatly, from innocent entertainment to intentional misdirection.
Political content
People may create deepfake videos of prominent political figures and/or candidates to potentially sway elections, alter public opinion, and mislead citizens.
Such deepfakes might be distributed across social media platforms and other online channels, appearing as legitimate political ads or official statements from the pictured politician.
Entertainment
Not all deepfakes are created and distributed for malicious reasons. Some are generated for entertainment purposes only, like making a celebrity say an absurd sentence or having a public figure sing your friend happy birthday. This content might be shared privately between individuals or shared on public platforms.
Forums
Users of online forums and public groups might come across deepfake content. Depending on the core focus or topic of the forum, the AI-generated photos and videos might be used to mislead group participants, share false information, or simply “troll” other users with outlandish content.
Tips for Spotting Deepfakes
With deepfakes becoming so widespread, how can you tell the difference between a real image and one that’s been created artificially?
Continued technological advancements have made this more challenging than ever, though there are still some common telltale signs that a photo, video, or audio clip is a deepfake.
Distortions or Inconsistencies
Deepfakes can certainly create believable spoof content. However, they are not foolproof, and errors in the photos and videos they generate are always possible.
AI will sometimes misrepresent key human characteristics and features, like adding an extra digit to their hand or distorting the face unnaturally.
Review images and videos closely to check for these errors if you suspect it’s a deepfake. If the appearance of the person in the content doesn’t align with what you’d expect from normal human behaviors and mannerisms, it might be a deepfake.
Natural Textures
On the other hand, sometimes you can tell the content is a deepfake because the image is too polished and smooth.
People, clothing items, and surroundings naturally have some uneven texture. However, AI-generated images often have an uncanny smooth appearance that looks cartoonish and fake.
For instance, does the texture of their skin match what you’d expect, given the person’s age? If they’re an older individual, do they appear to have the appropriate amount of wrinkles?
Audio Syncing
For video content specifically, check to make sure that the audio matches with the movements of the person’s lips.
While there can be lagging that leads to temporary mismatches, you should be able to tell if the person in the video is saying the words you’re hearing. Otherwise, it might be artificially generated.
Catching Deepfakes Can Be Difficult – Get Expert Support
Deepfakes also pose a significant challenge to businesses that must determine if their customers are legitimate. AuthenticID’s SVP of Global Solutions, Stephen Thwaits, noted in a recent article:
“Artificial intelligence (AI) has allowed bad actors to shift from low-volume fake IDs to deepfakes created solely for criminal purposes at a fraction of the cost of old-school fake IDs. With the high volume of personally identifiable information (PII) and biometric data available on the dark web or public social media accounts, AI engines can commingle genuine human information with AI-generated information and IDs. This is making the separation of good versus bad even more difficult. And unfortunately, deepfakes are currently here to stay.
In the identity verification world, the question of how you introduce a deepfake ID into a decisioning workflow isn’t a new one. Differentiating between a good customer uploading a genuine ID and a bad actor uploading a fake– and now a deep fake– is a core function.”
While the threat of deepfakes surges, AuthenticID continues to innovate to stop fraudsters. In 2024, our team introduced a new solution to detect deepfake and generative AI injection attacks.
Contact us to learn more about how your organization can meet the evolving challenge of fraud.