Police Remove Scientists from ADA Meeting: What It Means for AI Research

Five prominent scientists were physically removed from the American Diabetes Association annual meeting in New Orleans on June 5, 2026, for distributing reprints of an editorial critical of the Trump administration’s attacks on scientific research. The incident, first reported by MedPage Today and covered by Ars Technica, is a bellwether for the state of scientific discourse in the United States. For the AI research community, it is a direct warning.

The five scientists ejected were Steven Kahn, editor-in-chief of Diabetes Care and co-author of the editorial; former ADA President Desmond Schatz; Aaron Kelly of the University of Minnesota; Justin Ryder of Northwestern University; and Irl Hirsch of the University of Washington. They were handing out reprints of an editorial published in the ADA’s own journal, Diabetes Care, on April 29. The editorial sharply criticized the Trump administration’s cuts to biomedical research funding and its broader assault on the scientific enterprise.

Video footage from MedPage Today shows police confronting the researchers, attempting to snatch the editorials, and physically shoving Kelly. “They physically grabbed us, forced us out of the conference center, and now are telling us we can no longer attend this meeting,” Kelly told MedPage Today. “They’re taking our lanyards. It really has come to this in America. Censorship is real.”

The ADA defended the ejections, stating that the five violated the conference code of conduct, which prohibits “disorderly or disruptive conduct such as protesting.” The organization claimed the scientists were “respectfully given the opportunity to cease this behavior and chose not to.” Kahn was scheduled to chair a session and present at the conference. He was barred from doing so. Police warned the group that if they re-entered, they would be arrested.

The timing is critical. NIH Director Jay Bhattacharya had been scheduled to deliver the keynote address but pulled out at the last minute. A senior adviser spoke in his place. The editorial the scientists were distributing was published with a disclaimer added by ADA leadership insisting the organization had nothing to do with its content. Kahn told The Washington Post he planned to hand out 1,000 copies to encourage researchers “who feel their livelihoods are threatened by what the NIH is doing to science” to fight back.

This is not a story about diabetes. It is a story about the chilling effect on scientific dissent, and it directly implicates the AI research community.

The AI field operates under a similar, if less visible, pressure. The same administration that slashed NIH funding has also targeted academic AI research through visa restrictions, export controls on GPU access, and threats to revoke the tax-exempt status of universities that engage in certain research. The message is the same: toe the line or face consequences. The ADA incident makes the mechanism explicit. Dissent is not just discouraged. It is physically policed.

Consider the parallel. At the ADA meeting, scientists were removed for distributing a peer-reviewed editorial from the society’s own journal. The editorial was published. It was on the record. Yet the act of sharing it in physical form at the conference was deemed a code-of-conduct violation. What does that mean for an AI researcher who wants to distribute a preprint critical of a government AI safety policy at a conference where a government official is speaking? What does it mean for a graduate student who wants to hand out a critique of export controls on NVIDIA H100s at an industry event?

The answer is clear. The boundary between legitimate scientific discussion and “protest” is now determined by institutional discretion, not by the content of the speech. The ADA decided that handing out a journal reprint outside a room where an NIH official was speaking constituted disruptive conduct. The same logic could apply to any conference where a government representative appears.

The AI research ecosystem is particularly vulnerable because it is so deeply intertwined with government funding and policy. The Defense Advanced Research Projects Agency, the National Science Foundation, and the Department of Energy all fund large-scale AI research. The National AI Research Resource is a federal initiative. The AI Safety Institute is a federal body. If the ADA can eject scientists for distributing an editorial critical of the NIH, what happens when an AI researcher distributes a critique of the AISI’s evaluation methodology at a NeurIPS workshop where an AISI representative is speaking?

The answer is not hypothetical. The Trump administration has already shown willingness to punish academic institutions that host research it dislikes. In 2025, the administration threatened to revoke the tax-exempt status of universities that promote “divisive concepts.” In 2026, it moved to restrict visas for Chinese AI researchers. The ADA incident demonstrates that professional societies are willing to enforce these norms from within, without any government directive. The chilling effect is self-reinforcing.

There is a deeper structural issue. The ADA’s code of conduct prohibits “disorderly or disruptive conduct such as protesting.” But the definition of protest is entirely subjective. Handing out a journal reprint is not a protest unless the content of the reprint is deemed objectionable. The ADA’s action reveals that the code of conduct is a tool for content-based censorship, not a neutral rule about behavior. The scientists were not shouting. They were not blocking an entrance. They were handing out paper.

The AI community should take note. Conference codes of conduct are written to protect attendees from harassment. They are increasingly being used to suppress speech that conference organizers or sponsors find inconvenient. The ADA incident is the canary. The coal mine is every AI conference where a government official is scheduled to speak.

Kahn’s editorial, published in Diabetes Care, includes a passage that now reads as prescient: “It is no longer enough to stand idly by or work behind the scenes with lawmakers. Moreover, it is no longer appropriate to fret about political backlash. Now is the time to recognize and fight to reverse the spiraling fall of the United States of America’s status as the foremost nation in health care innovation.”

Replace “health care innovation” with “AI research” and the sentence holds. The AI community faces the same choice. Stand idly by while the space for dissent shrinks, or recognize that the fight for scientific freedom is a fight for the field’s future. The ADA ejected five scientists. The AI community should ask itself who it will eject next, and whether it will defend the right to distribute a journal reprint.