11,000 people took a test to classify 50 images as human or AI art. Median accuracy was 60%, indicating difficulty in differentiation. Many participants judged art based on style rather than content, leading to misclassifications. Notably, 60% of top-liked images were AI-generated. Those critical of AI art still preferred AI images. The results suggest that while AI art can deceive, human biases affect perceptions, reflecting broader feelings towards new art forms.
[https://www.astralcodexten.com/p/how-did-you-do-on-the-ai-art-turing](https://www.astralcodexten.com/p/how-did-you-do-on-the-ai-art-turing)