A recent study has revealed that training people to distinguish between real faces and those generated by artificial intelligence may be more effective than relying solely on automated detection programs. The results showed that participants were able to almost double their ability to detect fake faces after a short period of training, at a time when AI technologies are capable of producing highly realistic images that are difficult to differentiate from traditional methods.

Simple Training Achieves Remarkable Results

Researchers from the University of Aberdeen in collaboration with the Australian National University conducted an experiment to assess people's ability to distinguish between real faces and those produced by artificial intelligence. Instead of focusing on obvious visual errors, the training relied on observing more subtle features such as excessive symmetry in facial features, an unrealistically perfect appearance, weak emotional expressions, and difficulty remembering the face after looking at it.

The study's results showed a significant improvement in participants' performance, with success rates in detecting fake faces rising from around 40% before training to nearly 80% after a training session lasting about an hour. Some participants achieved accuracy rates close to 100%.

Artificial Intelligence is Evolving

The researchers believe that these results come at a time of increasing spread of deepfake technologies, which are being used in financial fraud, identity theft, and the dissemination of misinformation. This makes it even more important to develop human skills alongside improving automated detection tools.

As the study indicates, AI systems still struggle to produce accurate faces for certain demographics, such as children, the elderly, and some ethnic groups, due to biases in training data. This could provide indicators that help detect fake images.

The researchers conclude that the future of combating fake content will not rely solely on detection software but on combining technology with trained human capabilities. Humans will become an essential part of the system to counter digital forgery as AI technologies continue to evolve.