Your Country

7:00pm - Midnight

Now Playing: Toby Keith - A Little Less Talk (And A Lot More Action)

    SIGNS YOU’RE WATCHING A DEEPFAKE

    Matt Case - July 2nd, 2024 - 9:39am PDT 

    Election Year Heightens Deepfake Concerns: How to Spot Them

    NEW YORK — As the U.S., U.K., and France head to the polls in a crucial election year, social media is awash with concerns about disinformation, particularly deepfakes. These AI-generated images or audio of candidates are designed to mislead voters, causing widespread worry.

    So, how can you spot a deepfake? Experts recommend using common sense and skepticism, but also paying attention to these visual elements:

    1. Strange Looking Mouth or Chin: Oddities in this area, such as fewer wrinkles, less detail around the mouth, or a blurry chin, can be major giveaways. Additionally, if the voice and mouth aren’t synced, it’s a red flag.

    2. Weird Sentence Structure: Deepfake videos may feature unusual speech patterns, like saying “dollars” before a number. Comparing the clip to real recordings can reveal inconsistencies in voice, mannerisms, and expressions.

    3. Consistency Between Face and Body: Discrepancies in the size of the head relative to the body or differences in skin tone between the face and neck can indicate a deepfake.

    4. Lack of Continuity: Changes in clothing color within the same video should raise suspicion, especially if the person makes outrageous or unlikely claims.

    5. Extra Fingers, Hands, and Limbs: AI-generated images often depict abnormal numbers of fingers or oddly shaped hands, arms, or legs.

    6. Mangled Letters and Numbers: AI struggles to reproduce text and numbers accurately, resulting in illegible or nonsensical characters. Zooming in can help identify these flaws.

    7. Jerky Video Edits: Poorly manipulated videos, known as “cheapfakes,” are often easy to spot due to simple video-editing techniques.

    By staying vigilant and looking for these signs, voters can better protect themselves from falling victim to deepfake disinformation.