Can you Spot a Deepfake?

Deepfakes: Do you know one when you see one? Seeing is not necessarily believing these days. It’s not merely a matter of seeing a video of a celebrity saying something outrageous and then finding out later that it was a deepfake - cybercriminals are using them in scams -specifically phishing attacks. Let’s say you receive an urgent email that seemingly came from your boss, but there was something weird about it – something that just felt off.

Welcome to the world of deepfakes, enabled by a rapidly evolving technology that uses artificial intelligence (AI) to create synthetic media, often in the form of videos or audio recordings. They can appear real but are manipulated or completely manufactured.

This is not all bad, as there are legitimate purposes for deepfakes - people can use them for creative purposes like satire or entertainment, but their potential for misuse is both insidious and concerning. Deepfakes have already made it into political campaigns. In 2024, a fake robocall mimicked the voice of a candidate. Scammers wanted to fool people into believing they said something they never said.

Do you remember that news story a while back, where crooks produced a fake message to her mother that she had been kidnapped? They did that by obtaining a brief recording of the daughter’s voice. Bad actors can also use deepfakes to spread misinformation, damage reputations, and even manipulate financial markets.

The use of these fakes in phishing attacks is a major impact to business because they can provide employees with a stronger incentive to click on something they shouldn’t. Knowing how to identify different types of deepfakes is crucial in today’s world.

What are the different types of deep fakes?

Face-Swapping Deepfakes

Face fakes have already been around for a long time - they are the most common type. Here the face of one person is seamlessly superimposed onto another's body in a photograph or video. These can be quite convincing, especially with high-quality footage and sophisticated AI algorithms.

Here's how to spot them:

Consider the source: Where did you find the video” Was it on a reputable news site or a random social media page? Be cautious of unverified sources and unknown channels.

Look for inconsistencies: Check the overall lighting, skin tones, and facial expressions. Do they appear natural and consistent throughout the video? Look for subtle glitches such as hair not moving realistically or slight misalignments around the face and neck.

Listen up: Does the voice sound natural? Does it jibe with the person's typical speech patterns? Incongruences in voice tone, pitch, or accent can be giveaways.

Audio Deepfakes

Some scams start with trying to get you to leave a voice mail message somewhere so they can capture a recording. This type of fakie involves generating synthetic voice recordings by mimicking a specific person's speech patterns and intonations. Scammers can use these to create fake audio messages like making it seem like someone said something they didn't.

Here's how to spot those:

Focus on the audio quality: This type of fake audio can sound slightly robotic or unnatural, without the type of flow we’re used to. This especially comes into focus when compared to genuine recordings of the same person. Pay attention to unusual pauses, inconsistent pronunciation, and strange emphases.

Analyze the content: Does the content of the audio message align with what the person would say, or stay within the context in which it's presented? Consider if the content seems out of character or contradicts known facts about the ‘speaker.’ For example: Jokes from someone who never makes jokes.

Don’t trust without verification: Are any claims made be supported by any independent evidence? If not, approach it with healthy skepticism.

Text-Based Deepfakes

Text-based is a fairly new type of deepfake. It uses AI to generate written content like social media posts, articles, or emails. They mimic the writing style of a specific person or publication. These can be particularly dangerous because scammers can use these to spread misinformation or impersonate someone online.

Here's how to spot them:

Analyze the writing: If you focus on the writing style, vocabulary, and tone, you may spot anomalies. Does it match the way the person or publication typically writes? Look for unusual phrasing, grammatical errors, or inconsistencies in tone.

Is it factually accurate? Compare the information presented in the text against reliable sources. Don't rely solely on the content itself for confirmation.

Beware of emotional triggers: No matter the format, always be cautious of content that evokes strong emotions. This is an example of ‘social engineering,’ meant to evoke emotions such as fear, anger, or outrage. Scammers may be using these to manipulate your judgment.

Object Manipulation in Deepfake Videos 

This goes beyond faces and voices. Object manipulation uses AI to manipulate objects within real video footage - changing their appearance or behavior. Bad actors may be using this to fabricate events or alter visual evidence.

Here's how to spot them:

Physics and movement: How do the objects change within the video? If they move, does their motion appear natural and consistent with the laws of physics? Look for unnatural movement patterns as well as sudden changes in object size, or inconsistencies in lighting and shadows.

Find the original footage: If available, try to find the original source of the video footage. This can help you compare it to the manipulated version and identify alterations, which may be subtle.

Remain vigilant and apply critical thinking

This is the age of the deepfake. You should familiarize yourself with the different types of deepfakes and learn to recognize potential red flags. Verify information through reliable sources. These actions will help you become more informed and secure.