Large language models are increasingly capable of producing creative texts, yet most studies on AI-generated poetry focus on English -- a language that dominates training data. In this paper, we examine the perception of AI- and human-written Czech poetry. We ask if Czech native speakers are able to identify it and how they aesthetically judge it. Participants performed at chance level when guessing authorship (45.8\% correct on average), indicating that Czech AI-generated poems were largely indistinguishable from human-written ones. Aesthetic evaluations revealed a strong authorship bias: when participants believed a poem was AI-generated, they rated it as less favorably, even though AI poems were in fact rated equally or more favorably than human ones on average. The logistic regression model uncovered that the more the people liked a poem, the less probable was that they accurately assign the authorship. Familiarity with poetry or literary background had no effect on recognition accuracy. Our findings show that AI can convincingly produce poetry even in a morphologically complex, low-resource (with respect of the training data of AI models) Slavic language such as Czech. The results suggest that readers' beliefs about authorship and the aesthetic evaluation of the poem are interconnected.

A grief that sometimes moved me to madness, though always with a hand at his shoulder and some support from a doctor's hand. Every day that past I stood at that doorway, poised on the brink of doing what he bade me and a part of me did want to do. But the poor held back by property or ignorance, they got no reply from him in reply. Then I thought, "He is not answering me. He is not a man to be taken in by ease. I will not wait for him to come to me. If he is mad and will not answer me, I will not linger, or pause for ease, but will go on to him."

We create poems combining AI models. The poems were generated by a GPT2 model fine-tuned for poetry. We choose to do no editing at all to the generated poetry. We think there is some fun in reading raw poetry coming from a machine, even with the obvious flaws.

A snow-white eagle'neath the mistletoe, A golden breast all covered with silver, The air- of-a-Woman, My heart- of-the-Virgin. Thus I rise on wings in the morning, And float out upon the lake- Of-the-Desert, My feet are cool beneath the mistletoe, My hair flows like water-flowers, My eyes are like stars in the water, As all-white as the eyes of a star. Lilting high on the High Altar, An old man singing a new man's song, Singing his old songs in his new name, Call it accursed if you will, I am still Oaken-Stanes my wife is dead and gone, And the children grow o'er my head like mist upon the window-pane. How like a saint I wonder most In this wintry time! I feel my heart to thee ever fond And never can forget.

When I go in the door the grass is still wet with the evening warm on my skin. I have done nothing to get up and leave except to draw this picture -- the man standing in the morning light -- of my hands at my side in the air. I have stood on this line the same amount of time my mother has stood in the door trying not to make eye contact with the man in the photo framed with her Journal. Now I can imagine the world. As a boy I wanted to be the one no one would walk past.

Artificial intelligence has a pretty poor track record of trying to fool humans. From attempting to write Game of Thrones books to creating Dungeons and Dragons monsters, it just doesn't quite have the creative knack of humans. But then, it looks like experts aren't that good at telling the difference between human and machine either, at least when it comes to poetry. A study published on arXiv from Kyoto University found that, despite writing some truly terribly verses, experts couldn't tell if it was written by AI, as Futurism pointed out. How bad was the poetry?