People Can't Identify Fake AI Faces from Real Headshots

Dec 23, 2023 ·

AI technology has advanced to the point where it is difficult for people to tell the difference between a real and fake headshot. Recently, researchers at the University of Washington conducted an experiment that tested people’s ability to identify real headshots from faked AI-generated portraits.

The experiment was conducted using Amazon’s Mechanical Turk platform, which allowed the researchers to quickly gather a large pool of participants. The participants were asked to view pairs of headshots and determine which one was the real photograph and which was the AI-generated portrait.

The results showed that the participants could not differentiate between the real and fake images. In fact, the average accuracy was only 52%, meaning that nearly half of the time they were selecting the wrong image.

The researchers believe that their findings indicate that as technology advances, it may be increasingly difficult for people to distinguish between real and artificial images. This could have implications on the use of AI in fields such as law enforcement and security, where facial recognition is employed.

The researchers also noted that the difficulty in distinguishing between real and fake images could increase if deep learning algorithms are used to generate the fake images. The research team believes more work needs to be done to better understand how people process visual information when presented with AI-generated images.

Overall, this experiment shows that AI technology is advancing quickly and people may not be able to tell the difference between real and fake headshots. This could have implications for fields such as law enforcement and security, where accurate identification is essential. As AI technology continues to develop, further research will be necessary to ensure that people can accurately identify real and artificial images.