Hello Friend,
While AI may be great for productivity and helping us understand data sets quicker, it also helps create deepfakes: videos, images and audio that mimic real people.
While some deepfakes may be harmless fun, they are often used for scams, identity theft or cybersecurity risks. In this digital world, it’s important to know how to spot them to protect your business and yourself.
Here’s how you can detect a deepfake:
- There will likely be unnatural facial expressions, odd blinking rates and maybe even delayed lip-syncing.
- Look for inconsistencies in shadows and lighting, especially around the person’s face or body. If the lighting of the person doesn’t match the background, it’s a telltale sign that the video is a deepfake.
- See if there are any pixelated or blurry areas around the face like the hairline or ears. This will typically indicate AI tampering.
- Whether the deepfake is audio-visual or just audio, check out voice and lipsync mismatches, unnatural pauses or odd tone changes.
- Look for the lack of natural eye movement or too much eye contact, this can be an easy giveaway.
Is there an easier way to tell?
Yes, with AI getting better at being undetectable, tools have come out that will automate detection. These tools have a good accuracy rate, however we recommend double-checking their work.
