Deep fakes are taking over the Internet as artificial intelligence continues to appear more online.
Deep fakes are images or videos that someone distorts to convey a message different from its original purpose. This raises concerns as these photos contribute to the ongoing issue of distributing misinformation. The impact of the spread can cause mass confusion that have serious consequences. With campaign season underway, battling AI-generated deep fakes and misinformation poses a threat to voters for the upcoming 2024 presidential election.
Increase your awareness with these five tips to help recognize faulty content online.
Pay Close Attention to the Facial Features
One indication for spotting a deep fake is inconsistencies. The first place to look is on the subject’s face, specifically around the edges. Altering a person’s facial features is a common technique. If parts of the person’s skin tone do not match or if a portion of their face is too sharp or blurry, those are red flags that the image may be a deep fake.
Pay attention to see if their lips match the audio in the background. If their speech sounds manipulated or doesn’t align with their lip movements, someone may have tampered with the video.
Think About the Context
It is important to consider the context of content on websites and throughout social media. Hearing exaggerated commentary or seeing a questionable background that seems out of character for specific public figures could be AI-generated content. If you are still unsure about the validity of the context, try searching for the information from a trusted source.
A rule of thumb is to verify the information through other credible websites or other digital platforms. When dealing with a political figure or candidate, the information will likely be reported elsewhere. This makes it easier to determine the real content from the deep fakes.
Look For Misspelled Words
Humans make mistakes, but AI produces spelling errors, often gibberish, that are too simple for the average person to get wrong. In a recent TechCrunch article, the tech site reported that AI image generators perform poorly with objects that require more detail. Handwriting is a struggle for AI generators to construct. It is more common to find words misspelled when text is applied on a video or image.
Use AI Tools
An authenticator tool is one of the most useful methods for determining if a post is AI-generated. An authenticator tool analyzes a video or image to provide a score showing the amount of content manipulated. Microsoft is one of several tech companies that has the authenticator tool available to discern compromised material.
Watch Out For Strange Glitches
Glitches are sudden malfunctions that easily catch the attention of someone viewing the content. While glitches are common in technology, they look different when AI is involved. AI glitches distort certain parts of images and videos through pixelation. Pixels impact the quality of graphics by making them blurry while other areas of the image or video appears clear.