Experts warn of misinformation from advanced AI content

Generative artificial intelligence (AI) is becoming so advanced that it can be hard to tell which photos and videos are real. This is a concern because many people get their news from social media, making them vulnerable to misleading content. Experts like Siwei Lyu and Junfeng Yang emphasize the importance of skepticism when viewing online images or videos. AI-generated content can mislead viewers, especially in serious situations, like the recent wildfires in Los Angeles. Both experts suggest we should always question the authenticity of what we see. They warn that the signs used to detect AI content may soon change as technology improves. Here are some ways to spot AI-generated images or videos, though they aren't foolproof. For example, human anatomy is often not accurately represented. If you see people with unusual arm and hand positions, it could be a sign of AI content. Additionally, the movements of people in videos may appear unnatural. Lighting and shadows can also provide clues. If light is coming from one direction, watch if shadows fall consistently. AI-generated human faces often lack realistic eyes; pupils may appear irregular. If you closely examine AI-generated videos, you might see changes in facial features, such as teeth or lip movements. AI-generated voices currently lack natural breathing sounds, resulting in very clean audio. If you find a video too perfect or surreal, this may be a hint that it's AI-generated. Watch for unnatural repetition, like people or objects appearing multiple times in the background. When checking the content, always verify its source. Is it from a credible outlet? Look at descriptions or captions to see if they mention AI. It’s also important to check if the information aligns with trustworthy news reports to help you distinguish fact from AI-generated fiction.