How to Spot Deepfakes

By Dr. Jarrod Sadulski  |  08/24/2026

Why Scammers Use Deepfakes for Fraud

Artificial intelligence has brought a lot of change to our society. It is used in our daily lives – on smartphones and in smart assistants, for example – and makes our lives easier by providing easier access to information.

In business, artificial intelligence has increased our ability to go through massive amounts of data and find patterns that affect decision making. Artificial intelligence also impacts transportation, healthcare, and nearly every field that uses technology.

How Artificial Intelligence Creates Deepfakes

However, the use of realistic, AI-generated images is rapidly rising. These AI-generated images are created through machine learning models.

Through user prompts, AI software creates images by arranging pixels based on learned patterns. AI tools build an image based on a statistical assumption of how the image should appear.

But while these images can be used in positive ways, AI can also be used to make manipulated content such as deepfake images or videos. This content can be distributed online to misinform or embarrass people.

Common Techniques Used to Create Deepfakes

The creation of deepfakes involves generative AI and deep learning algorithms. A deepfake image or video may involve:

• Swapping a face
• Making other changes to someone’s appearance (such as facial hair transformations)
• Cloning someone’s voice and likeness

Deepfakes are dangerous, however. A deepfake can create the illusion that someone stated something that they didn’t actually say in real life, which can lead to all kinds of trouble.

High-end deepfake manipulations can be used for many nefarious purposes, such as:

• Reputational damage
• Identity theft
• Crimes against children
• Online grooming for human trafficking
• Disinformation
• Extortion

Deepfake videos can be used to spread propaganda and fake news by altering what a public figure – such as a politician or a community leader – states. Celebrities’ images and voices have also been cloned and used to create convincing scams, according to Brooke Seifel of McAfee®.

Deepfake technology can have an adverse impact on the criminal justice system. When criminals use deepfakes, it is more difficult for investigators and the court system to determine what evidence is real and what is fabricated.

Detecting AI-Generated Images and Deepfake Videos

As technology advances, it becomes harder and harder to spot a deepfake. However, it is possible if you look closely.

First, it is essential to review the source. For instance, is that source capable of creating deepfakes? Does the source have something to gain by showing a deepfake image or video to you and upsetting you?

Second, look carefully at a video’s visual effects. Indicators of a deepfake include:

• Irregular lip movements that aren’t in sync with a verbal message
• A robotic tone to a person’s speech
• Unusual patterns of a person blinking or no blinking at all
• Unusual facial reactions that don’t match what is said
• Unnatural eyebrow movement
• Odd body mannerisms
• Inconsistent body movements that don't match how a real person moves

Also, deepfake technology struggles with lighting and proper placement of shadows. Additional indicators of a deepfake image or video involve:

• Too much glare on someone’s skin or glasses
• Facial transformations or face swapping
• Unnatural glare changes
• Shadows that are out of alignment in a person’s hair
• No match in skin tone between a person’s forehead and neck
• Blurring or flickering during scene transitions in a video

Detecting Audio Deepfakes

Audio deepfakes can be more challenging to detect than video or image deepfakes and there is not always a single, telltale sign of fraud. Some indicators that a deepfake has been created through AI-generated voice manipulation include listening carefully to see if there are:

• Unnatural pauses when a person speaks
• Vocal inflections that don’t match what is being said or expressing unnatural emotions that ordinary people don’t use
• Unrealistic statements
• Slurred or unnatural speech
• Odd background noises like static or crackling, designed to mask voice cloning problems

Scams Involving Dangerous Deepfakes

High-quality deepfakes can be convincing and unfortunately, they are good enough to be commonly used in scams. For example, a scammer can clone the voice of a scam victim’s family member and have that “family member” ask for help due to an emergency. The victim is then convinced to part with money or sensitive information such as a credit card number.

There have been cases where a scammer targets someone and says that a child or grandchild is being held hostage or has been arrested. Due to the realistic-sounding audio, this scam can be believable if the fake voice sounds identical to the loved one.

Warning signs of a deepfake scam include:

• The speaker conveys the need to act urgently or maintain secrecy
• The timing of the communication is suspicious
• The speaker makes statements that are out of character
• An unusual communication pattern

To verify if a text, phone call, image, or video is legitimate, it is best to confirm its validity through another source. For instance, call or text the cell phone of the “victim” to see if he or she answers before taking action.

Deepfakes are commonly associated with anonymous accounts and newly created accounts, so it is always important to consider information sources and pay attention to anything that doesn't seem right.

Deepfake Detection Software and Systems

Antivirus software is a useful way to detect deepfakes. For example, antivirus software can conduct scans in real time and identify if an image, voice message, or video isn't real.

In addition, automated detection systems can be used on social media sites and other online platforms to detect deepfakes.

Why Deepfakes Are Becoming More Common

Deepfakes have become an indelible part of the digital world. As technology continues to evolve, we can expect scams involving deepfakes to become more common. It is important to remain up to date on current trends involving AI and deepfakes, which can help you to avoid becoming a victim of scams or fraud.

In the future, deepfakes will likely become more complex, more interactive, and harder to detect, which has serious implications regarding safety, security, and legal risks. More laws are needed to protect victims from problems involving AI and deepfakes. Without more laws, deepfakes can increase the risks for fraud, political propaganda, and social engineering.

The Bachelor of Science in Criminal Justice at APU

For adult learners interested in studying different aspects of criminal justice, American Public University (APU) offers an online Bachelor of Science in Criminal Justice. This degree program offers courses such as criminal profiling, criminalistics, and constitutional law. Other courses include crime analysis, criminal law, and criminal investigation.

This B.S. in criminal justice offers a concentration in digital forensics. This concentration offers courses in cybercrime, wireless networks, computer forensics, and various aspects of digital forensics.

For more details, visit APU’s security and global studies degree program page.

Note: This degree program is not designed to meet the educational requirements for professional licensure or certification in any country, state, province, or other jurisdiction. This program has not been approved by any state professional licensing body and does not lead to any state-issued professional licensure.

McAfee® is a registered trademark of McAfee, LLC.