Astonishingly Prolific Anti-Israel Medical Researcher Used AI to Churn Out Journal Articles, Screening Software Shows

A stridently anti-Israel Harvard graduate student used artificial intelligence to churn out at least five of the more than 90 medical journal articles he published in two and a half years, including one about whether newborns have a future in Gaza, a Washington Free Beacon review shows.

“The ongoing Israeli military assault on Gaza has led to an alarming humanitarian catastrophe, whereby the onset of famine is coupled with a deterioration of maternal health services, severely impacting the wellbeing of pregnant women and of children. The near-total collapse of the health-care infrastructure, coupled with the lack of access to essential medical services, has resulted in a tragic surge in preventable maternal and neonatal deaths,” said the article titled “Will there be a future for newborns in Gaza?” in the November 2, 2024, issue of the Lancet, a British medical journal. “The world cannot remain silent any longer. The time for action is now—to restore access to health care, to protect women and children, and to uphold the sanctity of life.”

The lead author of the article, Bilal Irfan, published a Harvard Medical School email address for correspondence related to the article, and he listed his affiliation as “Center for Bioethics, Harvard Medical School, Boston, MA 02115, USA.” At the time the article was published, he was a graduate student at Harvard Medical School. The author listed second on the article, Abdallah Abu Shammala, listed his affiliation as Gaza’s European Hospital. The Israel Defense Forces has posted video of Hamas terrorist tunnels under that hospital, and in May 2025 the IDF said it had killed “the terrorist Mohammed Sinwar, Head of the Hamas terror organization’s military wing,” along with two other terrorists, “in an underground command and control center, under the European Hospital in Khan Yunis.”

Three online screening programs—Pangram, Winston AI, and ZeroGPT—marked the article as 100 percent AI-generated. Another program, Quillbot, said 54 percent of the text of the article is AI-generated. A fifth program, GPTZero, gave the article a 73 percent likelihood of being AI-generated.