Harmful use of AI-generated Content - A Study on the Impact of Cognitive Ability on Identifying AI-Generated Images

Sammanfattning

Advancements in generative artificial intelligence have led to an increase in the usage of such tools, both by individuals and organizations. While it has been helpful to make organizations more effective, it can also be used in harmful ways. There has been an emergence of AI-generated content to spread convincing misinformation, also known as deepfakes. Deciphering what is real and what is not is considered to be based on a multitude of factors such as attention, cognitive ability, beliefs, emotions, and liking patterns. This study investigates to what extent the cognitive ability of the individual affects whether humans can identify AI-generated images. This is done through a quantitative study based on internet surveys where both the cognitive ability and the ability to identify deepfakes are measured for each participant. The data is then analyzed to see if there is any correlation between the two. In contrast to previous research, the results from this study indicate that there is little to no correlation between cognitive ability and the ability to identify AI-generated images. The study also suggests no correlation between the difficulty level of the correctly identified images and the cognitive ability of the participants.