University of Hong Kong professor steps down from associate deanship after AI-generated references scandal

A University of Hong Kong (HKU) professor has stepped down from his position as associate dean after a paper that listed him as a corresponding author was found to contain non-existent references generated by artificial intelligence (AI).

Professor Paul Yip, director of the Hong Kong Jockey Club Centre for Suicide Research and Prevention at the University of Hong Kong, meets the press on September 10, 2025. Photo: Kelly Ho/HKFP.

HKU announced on Wednesday that Professor Paul Yip of the Faculty of Social Sciences had resigned from his position as associate dean, as part of a series of disciplinary actions and remedial measures following the AI citation scandal that broke last month.

The scholar also withdrew his membership from the faculty’s research committees, the university said. He remains director of the Hong Kong Jockey Club Centre for Suicide Research and Prevention at HKU.

Also on Wednesday, HKU announced that its investigation into the paper “Forty years of fertility transition in Hong Kong,” a doctoral thesis published in China Population and Development Studies, found that some listed references were non-existent publications generated by AI.

The technology’s use was not disclosed by the primary author, PhD student Bai Yiming, while Yip was listed as the corresponding author of the paper, the university said.

Yip has requested that the paper be retracted from the journal, and the doctoral student is now the subject of disciplinary procedures, HKU said.

The University of Hong Kong. File photo: GovHK.

“HKU reiterates its commitment to rigorous academic and research standards. All researchers at HKU are required to ensure their work meets internationally recognised benchmarks for quality and ethical conduct,” the university said on Wednesday, repeating what it stated last month.

The university, which last month reclaimed its position as Asia’s top university in the 2026 QS Asia University Rankings for the first time in 15 years, also vowed to strengthen training for all researchers regarding the use of AI in academic research. Training and assessment in such areas is “mandatory” to ensure the preservation of academic integrity, the university said.

Paper retracted

According to Springer Nature, the publishing company of the academic journal China Population and Development Studies, it launched an investigation on November 15 after concerns about the reference accuracy in the article were raised.

The paper was officially retracted by the journal’s editor-in-chief on Monday.

“Upon further investigation, the Publisher was unable to verify the source of at least 24 references. The authors admitted the use of AI tools without declaration,” Springer Nature said in a statement.

“The editorial office acknowledged an oversight in verifying the authenticity of the references during the editorial process. The Editor-in-Chief therefore no longer has confidence in the reliability of this article. The authors agree with this retraction.”

A retraction note for a doctoral thesis by University of Hong Kong PhD student Bai Yiming, titled “Forty years of fertility transition in Hong Kong,” published in the academic journal China Population and Development Studies. Photo: Screenshot, via Springer Nature.

The suspicions about the paper first came to light on Threads in early November, with a netizen claiming that most of the sources cited in the paper were “non-existent.”

Local newspaper Ming Pao reported at the time that among the 61 references listed – which included academic articles, books, government statistics, and Legislative Council documents – 21 items contained a Digital Object Identifier (DOI) link, while 35 included links to Google Scholar.

However, upon clicking the links, seven DOIs were listed as “not found,” and 22 Google Scholar links showed that the articles could not be found, the report read.

Yip told local media outlet HK01 last month that the student used AI to “tidy up” the references but had not checked the content.

He admitted responsibility as the corresponding author of the paper and apologised for the incident. The professor had defended the academic integrity of the authors and said the paper was not fabricated and had passed two rounds of review.