AI-Generated Fakery: A Growing Threat Online

Deceptive photos, videos, and audio are spreading rapidly due to powerful new AI tools. These "deepfakes," featuring everyone from celebrities to politicians, make it increasingly difficult to distinguish truth from fiction.

Tools like DALL-E, Midjourney, and OpenAI's Sora allow anyone to create deepfakes with ease. While seemingly harmless, these fakes can be used for scams, identity theft, propaganda, and even election manipulation.

How to Spot a Deepfake

The early days of deepfakes were a bonanza for eagle-eyed viewers. Glitches like extra fingers or mismatched glasses were clear giveaways. However, AI advancements have made detection trickier. Common advice, like looking for unnatural blinking, may no longer be foolproof.

Here are some things to watch for, but be aware that skilled deepfake creators can sometimes overcome these tells:

• Unnatural Smoothness: Deepfake photos, especially portraits, may have an overly-polished, almost electronic appearance to the skin. Look for a lack of natural imperfections or subtle variations in skin texture.
• Lighting Inconsistencies: The person in the deepfake might be in sharp focus with realistic lighting, but the background may appear less detailed or have inconsistencies in lighting and shadows. Look for unnatural shadows around the face and hair, or inconsistencies in light source direction between the subject and the background.
• Uncanny Movements: While deepfakes of facial expressions have become more sophisticated, pay attention to subtler movements. Does the person blink at a normal rate and with natural variation? Do their head movements appear fluid and natural, or are they jerky or stiff?

Beyond these visual cues, consider the content itself. Is the video or image out of character for the person featured? Does it seem too good to be true, or promote an outrageous claim?

By employing a critical eye and maintaining a healthy dose of skepticism, you can become more adept at spotting deepfakes and protecting yourself from online misinformation.

Deepfake Deception: How to be a Media Sherlock

The age of information overload can be a breeding ground for misinformation. But fear not, there are ways to be a savvy media detective! Deepfakes, those cleverly manipulated videos, can be tough to spot, but with a keen eye and a dash of skepticism, you can become a truth-seeker.

Facial Forensics:

• Edges Under Scrutiny: Deepfakes often stumble on the finer details. Examine the person's face along the hairline and jawline. Does their skin tone seamlessly blend with their neck and ears? Are the edges crisp, or are they blurred and unnatural?
• Lip Logic: If the person is talking, pay close attention to their mouth. Do their lips move fluidly and match the words being spoken? Look for unnatural pauses or inconsistencies between speech and lip movement.
• Tooth Trouble: Deepfakes can struggle with intricate details like teeth. Are the teeth clear and distinct, or do they appear blurry or oddly uniform? Some experts suggest that a lack of definition around individual teeth might be a sign that something's fishy.

Beyond the Surface:

• Context is King: Don't just watch, analyze! Does the content align with what you know about the person or situation? Is what you're seeing believable? Consider the source and any potential motives behind the video.
• Reverse Image Search: If you're suspicious, run a web search for the video or screenshots from it. You might find the original footage or debunking articles from credible sources.

Remember, deepfakes are constantly evolving, so staying informed is key. By being a critical thinker and applying these techniques, you can become a master of media discernment and expose deepfakes for what they are: elaborate illusions.