BOZEMAN — With a growing number of artificial intelligence (AI) platforms accessible to the public, it’s becoming harder to tell what’s real and what’s generated by artificial intelligence.
To learn how to tell what’s real and what’s AI, I headed to Montana State University and talked with Matthew Revelle, a computer science professor.
Cassidy Powers reports - watch the video here:
Revelle has taught computer science at MSU for three years and has been interested in the subject his entire life — starting as a kid who loved video games, ultimately earning his Ph.D. from George Mason University. He explained that generative AI can create original content, such as text, images, music, and code.
“At a glance, it can be really difficult to determine whether something is real,” Revelle said. “An important first step is if you see something that looks strange.”
Revelle demonstrated examples of AI videos found on social media.
Looking at one clip, he noted, “I would say the physics. Something looks slightly off.”
In another video, he pointed to shadows and lighting as potential giveaways. Other indicators of AI content, he said, include blurring, glowing, and visual inconsistencies.
When asked about the effectiveness of AI-detection websites, Revelle cautioned that because AI technology is advancing so rapidly, “It’s more difficult to do the detection than it is to do the generation.”
His advice on AI tools becoming more common? “Be skeptical. Just like when we read some article on the internet that makes a claim with little to no evidence, fact-checking is important.”
