Can You Recognize a Deepfake? Here’s What You Need to Know!
Let’s discuss what deepfakes are, recognize common scams, and learning detection techniques, you can stay one step ahead.
So what are deepfakes?
Deepfakes are AI-generated videos, sunds or pictures that look real which makes someone say or do things they never did.
You’ve probably seen some of these recently in the news. For instance, a 53-year-old French woman, known as Anne, believed she was in a romantic relationship with Brad Pitt after receiving messages from scammers posing as his mother and then him. She divorced her wealthy husband, sent $9,000 for “customs fees” on gifts, and ultimately lost over $800,000 from her divorce settlement [1].
In another case, a retired lecturer from Edinburgh named Nikki MacLeod was duped out of £17,000 in a romance scam involving deepfake videos of a woman she thought she was dating [2].
The Ugly Side of Deepfakes: Common Scams
Here are common types of deepfake scams:
1. Investment scams
Scammers use deepfakes to impersonate trusted figures like celebrities or executives promoting phony high-return opportunities. Take 82-year-old retiree Steve Beauchamp, who saw a video of Elon Musk touting a cryptocurrency investment. Believing it was real, he invested over $690,000 from his retirement savings, only for the scammers to vanis
These fakes prey on greed and trust, leading to massive financial losses.
2. Romance Scams
You’ve probably heard of people meeting someone online and getting caught up in a whirlwind romantic courtship, only to discover their love interest is a con artist. As in the Brad Pitt case mentioned earlier, scammers exploit loneliness, convincing people to part with life savings or even end marriages. The Edinburgh woman’s story highlights how AI videos can make these cons feel heartbreakingly authentic.
3. Polictical scams:
Deepfakes can smear reputations, spread lies, and incite chaos. Bogus clips have gone viral, influencing elections or tanking stocks.
In Slovakia, a fake audio of a political leader discussing vote rigging surfaced just before the 2023 parliamentary elections, potentially swaying public opinion. These tactics erode trust in institutions and can spark real-world unrest.
4. Extortion scams
With this type of depefake scam, scammers mimic loved ones or authority figures to demand urgent payments. A chilling example: A finance worker at a Hong Kong multinational firm transferred $25 million after a video call with deepfake versions of the company’s CFO and colleagues, who instructed the transfer for a “secret transaction”. This breeds paranoia, as even familiar voices or faces become suspect.
A Simple guide to spotting deepfakes:
There are several telltale signs that humans can look for when discerning real from fake images, including the following:
In Videos:
- Unnatural body posture or facial movements, like odd blinking patterns.
- Inconsistent reflections (e.g., in eyes or glasses) or shadows.
- Audio out of sync with mouth movements.
- Blurred areas around the mouth or chin.
- Excessive pixelation or jagged edges.
- Discontinuities, such as a sudden change in clothing color.
In Photos:
A glossy, “too perfect” appearance.
Features that seem off, like distorted limbs, extra fingers, or unnatural facial structures.
Warped or illegible text and numbers.
Mismatched skin tones between face and body.
In Audio:
Robotic or flat tones lacking natural variations.
Absence of background noise or unnatural-sounding ambiance.
Tools like Sensity AI, Reality Defender, Truepic Lens, and DuckDuckGoose can help automate detection, but starting with these manual checks is a solid first line of defense.
The bottom line
Remember to keep a sharp eye and a healthy dose of skepticism, you can often spot the signs of a fake. Look out for unnatural movements, inconsistent details, distorted visuals, or robotic voices.
While tools like Sensity AI, Reality Defender, Truepic Lens, and DuckDuckGoose can support detection, your own vigilance is the first and most important defense.
Don’t be rushed and don’t be hushed.
References
[1] https://www.newsweek.com/brad-pitt-ai-dating-scam-france-deepfake-social-media-2014546
[2] https://www.bbc.co.uk/news/articles/cdr0g1em52go
[3] https://www.media.mit.edu/projects/detect-fakes/overview/
[5] https://www.ncoa.org/article/understanding-deepfakes-what-older-adults-need-to-know/
[6] https://www.techtarget.com/searchsecurity/tip/How-to-prevent-deepfakes-in-the-era-of-generative-AI
