Chinese Officials Rescind Calligraphy Prize Over AI-Generated Entry

The emergency management office in Dongguan, China, rescinded a second-place prize awarded in a calligraphy contest after internet users demonstrated that the winning entry had been entirely generated by artificial intelligence. The case, which unfolded over three days in mid-July, sparked a broad discussion in China about how art contests should verify submissions in the age of generative AI.

How the Fraud Was Detected

The contest was organized by the local emergency management office in Dongguan, a city in Guangdong province. The entry aimed to promote safety awareness through a classical Chinese poem written in calligraphic script, and it took second place in the calligraphy and literature category. The results of the contest's second round were published on the office's official WeChat account on July 12.

Two days later, internet users began publicly questioning the authenticity of the winning entry. They pointed out that every character was identical in size and height, the spacing between characters was rigidly uniform, and the writing lacked the natural traces of handmade calligraphy, such as varying brushstroke thickness or uneven ink saturation. The punctuation marks also raised suspicion, as they looked identical throughout the entire piece, something practically impossible to achieve by hand.

Official Response

On July 15, after an internal review, the office confirmed the users' suspicions and announced that the work had indeed been created using artificial intelligence. The entry was immediately removed from the list of winners, and the organizers said that those responsible for allowing the submission through would be dealt with in accordance with applicable regulations. According to Chinese media, the evaluation process focused mainly on the substantive quality of the content and failed to check whether the work had been created by hand.

The work was clearly identified as AI-generated content and immediately withdrawn from the list of prize winners - statement from the Dongguan contest organizers

Not the First Case

The Dongguan case is not an isolated incident. China was earlier rocked by controversy over the Hohhot Photography Bimonthly contest, in which first prize was awarded to a photo depicting three public order officers sitting on a bench. That image also turned out to be AI-generated, revealed by details such as distorted text on the officers' uniforms and unnatural hand positioning. The local literature and arts association subsequently rescinded the prize.

Both cases follow the same pattern: AI-generated works can slip past juries focused on content, only to be exposed later by internet users spotting small visual inconsistencies that generative models still cannot fully hide. In calligraphy, these are the rigid regularity of the characters and the absence of any trace of a physical writing instrument; in photography, it's distorted text or unnatural hand anatomy.

Implications for Cultural Contests

Commentators cited by Chinese media, including representatives of the cultural sector, say the growing number of such scandals is forcing organizers of literary, photography and art contests to introduce formal procedures for verifying the origin of submissions, going beyond simple quality assessment. Without such mechanisms, categories reserved for human creativity become vulnerable to machine-generated submissions, which can push the work of genuine authors out of prizes and funding.

For Polish readers, the case is a warning sign for domestic cultural, literary and photography contests as well, which still rarely have clearly defined procedures for detecting and disqualifying AI-generated entries. In Poland, similar controversies have so far been mostly limited to debates over labeling AI-generated advertising and journalistic content rather than art contests, but the experience of Chinese institutions shows the problem can arise in any field that relies on juror assessment of submissions.

Dongguan authorities did not say whether the author of the submission would face any consequences beyond losing the prize, nor whether their identity has been disclosed. It also remains unclear whether the contest will be rerun in the disputed category or whether the office will introduce new verification rules for future editions.