What are Deepfakes and Cheapfakes, and how can we spot them?
What are Deepfakes and Cheapfakes, and how can we spot them?
Embedded Files
tl;dr:
Deepfakes and Cheapfakes are everywhere and spotting them is going to get harder and harder as technology advances. While being able to identify them is important, what matters is looking at the intent behind the content.
Is it trying to manipulate you? Is it designed to trigger an emotional reaction? Even real images and videos can be used to mislead or pressure you into a certain belief or action. Before reacting or sharing, take a step back and ask yourself who benefits from you believing it or spreading it!
What are Deepfakes and Cheapfakes?
A deepfake is an AI-generated fake that manipulates someone’s face, voice, or actions in a way that looks and sounds real. AI models create these to make it seem like someone said or did something they never actually did. A deepfake could show a politician giving a speech they never made, a CEO issuing fake instructions in a video call, or a celebrity in a fabricated interview.
A cheap fake is a low-tech version that does not rely on AI. Instead, simple editing tricks like slowing down, speeding up, cutting, or mislabeling videos are used to push a false narrative. A common example is a video of a public figure slowed down to make them appear confused or intoxicated or a real event taken out of context to mislead people.
Why would someone want to do this?
People use deepfakes and cheap fakes to manipulate others because they are powerful tools for control, influence, and deception. The goal is not just to trick you into believing something fake, but to push you toward a specific reaction such as fear, outrage, trust, or even inaction.
In a corporate space, our biggest risk is these being used for Social Engineering attacks to convince someone to give them access to something or commit some other Scam or Fraud. Attackers can use AI-generated voices or videos to impersonate executives, family members, or trusted figures to trick people into sending money or revealing sensitive information.
If someone wants to pressure or manipulate you, they do not need the content to be real, just believable enough. A fake video, an out-of-context clip, or a doctored image can be enough to make you act without questioning.
Some other reasons could be:
Misinformation and propaganda – Fake content can shape public opinion, sway elections, or turn people against each other. A deepfake of a politician saying something outrageous can spread quickly before anyone verifies it.
Character attacks and blackmail – False media can be used to ruin reputations, discredit whistleblowers, or silence people. The more realistic it looks, the harder it is to prove it is fake.
Distraction and confusion – If people flood the internet with deepfakes, it becomes harder to know what is real at all. The goal is to create doubt so that even real evidence can be dismissed as fake.
We see many of these attacks being done on Social Media. Remember that real content can be used to manipulate you just as easily as AI generated content, so always be aware of this and as with any other form of Social Engineering, be Politely Paranoid! If content you see causes any kind of emotional reaction, take a pause, check the facts, and think twice before forwarding, commenting etc.
How to Spot Fake or AI Generated Photos and Videos
The technology used to create fake content is improving all the time, so these tips will become less valuable as time progresses. For now, look out for the following things which MAY reveal the origin of content:
Too shocking to be real - Does the image show a celebrity, politician, or cultural figure doing something completely unexpected?
Weird body parts and clothing - Are hands, fingers, or limbs strangely merged? Do background people have missing faces? Are earrings mismatched or embedded in skin, like on a forehead?
Overly smooth and overly colorful - Does the image have an unnatural glow, with lighting that seems to come from all directions? Do the colors look too intense or unrealistic?
Designed to manipulate emotions - Is it a polished image of a child or soldier holding a sign asking for money, likes, or shares? Look for strange misspellings in these types of images.
Glitchy mouth and teeth - Do the lips and teeth fall out of sync? Do you notice brief flashes of oversized teeth when the mouth is closed?
Lip sync issues - Do the words perfectly match the mouth movements, or is something slightly off?
Unnatural blinking - Does the person blink too slowly, too often, or in a way that looks like it is in reverse?
Odd breathing patterns - Does the person go too long without a breath, then suddenly take exaggerated inhales?
How to Spot AI Voice Clones in Phone Calls or Robocalls
Overly talkative - AI-generated voices, especially those connected to language models, tend to speak for much longer than a human would.
Long, awkward pauses - AI struggles with real-time responses. If you interrupt, there will either be a noticeable delay or an unnatural-sounding recovery.
Scam calls or messages pretending to be your boss - If someone claiming to be your CEO, CFO or anyone else senior in the company urgently tells you to transfer money, verify their identity and the real request before taking any action.
Deepfake technology is advancing quickly. Staying informed and sharing what you learn will help protect your coworkers and loved ones.
Page updated
Report abuse