AI Health

Friday Roundup

The AI Health Friday Roundup highlights the week’s news and publications related to artificial intelligence, data science, public health, and clinical research.

June 5, 2026

In this week’s Duke AI Health Friday Roundup: study tracks the spread of LLM influence through the literature; China approves brain/chip interface for commercial use; automation, datafication, and mutual aid in healthcare; applying public values to AI governance; DEET may fail us now; AIs: engines of efficiency or generators of additional burdens?; survey tracks adolescent use of chatbots; much more:

AI, STATISTICS & DATA SCIENCE

“An analysis of 7.3 million journal articles from 2020 to 2025 reveals that by 2025, slightly over half show evidence of LLM influence, with usage varying markedly by world region, institutional prestige, publisher, and discipline. These patterns demonstrate that AI adoption in science is not uniform but shaped by social and institutional forces including language barriers, competitive pressures, and editorial gatekeeping.” A research article published in PNAS by Kyle Siler traces evidence of the increasing diffusion of LLM-generated text in the published scholarly literature over the last few years.
“This emerging technology requires more than abstract lists of ethical principles or positions that connect at the lowest common denominator. As policy scholars and moral philosophers cast around for consensus principles, they should explore the fact that every society already operates on a set of near-consensus beliefs—what policy scholar Barry Bozeman calls public values.” In the spring edition of Issues in Science and Technology, John Nelson calls for the incorporation of “public values” into deliberation and debate about ethical application of AI technologies.
“In this nationally representative survey study of US adolescents and young adults, a fifth reported using AI chatbots for mental health advice. AI chatbots are already embedded in many youths’ mental health information ecosystem, underscoring the need for parents and clinicians to proactively discuss chatbot use to promote safety, appropriate expectations, and linkages to evidence-based care.” A national survey study published in JAMA Pediatrics by McBain and colleagues provides a picture of current AI chatbot use by adolescents and young adults in the US.
“Mathematicians have a choice about whether and how to adopt artificial intelligence in the conduct of their research. They also have a responsibility to ensure the continued flourishing of the discipline. This Declaration calls upon mathematicians to exercise this responsibility, and provides recommendations for individuals, institutions, government, and industry.” A group of academic mathematicians have joined a declaration of principles on the use of AI in mathematical research.
“Workers in some sectors have found major time savings from AI. But for others, the tools just change the work rather than making it faster. Workers might be spending less time writing patient portal messages, for example, but more time editing the releases the AI tool writes…At best, this mismatch between employer expectations and employee reality can be an annoyance. In other cases, however, it can result in workers being laid off for failing to meet unrealistic efficiency demands.” Vox’s Anna North examines the evidence for whether AI applications are actually making workers more efficient – or merely moving the burden further downstream.
“In April, GitHub announced that it was moving subscribers from request-based billing to a usage-based model for its AI-powered Copilot service. As that new pricing model goes into effect today, many GitHub Copilot users are reporting some extreme sticker shock as they realize just how quickly their previous “normal” usage is burning through their newly limited monthly allotment of AI credits.” Ars Technica’s Kyle Orland reports on the “sticker shock” experience by some AI users as pricing models change.

BASIC SCIENCE, CLINICAL RESEARCH & PUBLIC HEALTH

“…research published yesterday in the Journal of Experimental Biology suggests that mosquitoes may learn to associate the smell of DEET with dinner—and start gravitating toward it instead of away from it. The findings challenge long-held assumptions about how DEET works and what mosquitoes may be capable of learning….When the trained mosquitos caught a whiff of DEET alone, more than 60% of them tried to feed again…” Hey, everybody! CIDRAP’s Laine Bergeson has some GREAT news about mosquitoes for the summer season!
“During a procedure that took just over an hour and a half, the device’s sensors, which collect Dong’s brain signals, were placed on his dura mater, the tough outer layer of tissue that covers and protects the brain. The signals are transmitted to a computer by an implant placed on Dong’s skull. The computer then translates the signals into commands for a soft robotic glove Dong wears during the 2.5-hour training sessions he completes each day to help him learn to grab.” MIT Technology Review’s You Xiaoying relays the account of China’s first commercially approved invasive brain-computer chip interface in a human participant.
“In this trial, in which 91.8% of the patients had RAS G12 mutations, treatment with daraxonrasib led to significantly longer overall survival and progression-free survival than the investigator’s choice of chemotherapy in the RAS G12 and overall populations of patients with previously treated mPDAC, and the percentage of patients with a response was twice as high with daraxonrasib.” A research article published in the New England Journal of Medicine by O’Reilly and colleagues reports on a clinical trial that pitted a new therapy, daraxonrasib, against standard chemotherapy for previously treated metastatic pancreatic ductal adenocarcinoma.
“…we demonstrate that explanted epidermal, connective, neural, and muscle tissue from the sea cucumber Psolus fabricii (Holothuroidea: Echinodermata) healed and continued to grow in natural, nonaxenic seawater without supplementation for more than 3 years…. Our findings challenge conventional perceptions of tissue immortality and present a new class of experimental model, free from ethical concerns, with substantial implications for regenerative biology, biomedical research, and tissue engineering.” A paper published in Science Advances by Jobson and colleagues reports on the remarkable properties of a variety of sea cucumber whose explanted tissue appears to be immortal under ordinary conditions.

COMMUNICATIONS & Policy

“Intensified datafication and automation also make healthcare more abstract and opaque, in turn making it difficult for patients to navigate these systems, or to participate meaningfully in their governance. Decisions about healthcare delivery and processes are increasingly centralized in the hands of a small set of professionals with expertise in extremely complex incentive structures, payment schedules, and the information and computational systems that undergird them.” An essay by Linda Huber for Data & Society reflects on the possibilities of digital mutual aid as a counterpoise to the ongoing centralization and “datafication” of healthcare.
“The Erlanger case, which has not been previously reported, offers a rare glimpse at an apparent failure of AI drug diversion software used in hundreds of U.S. hospitals with little transparency or oversight. Healthcare facilities are not required to disclose their implementation of this kind of software or report malfunctions to anyone, so there is no full account of how widely these programs are used or how often they fail.” An article by Brett Kelman and Darius Tahir at Kaiser Family Foundation Health News reports on the failure of an AI system that was supposed to alert hospital employees to signs that opioid drugs were being diverted from patients.
“The present study is also one of the first studies that examines the actual behavior of victims rather than relying on hypothetical or staged responses, and the first known study to assess whether a CM [counter-marketing] strategy can be effective for preventing mail fraud victimization. We find that the CM approach reduced fraud revictimization by 17.9% relative to no intervention.” A research article by Langton and colleagues, published in the Journal of Consumer Affairs, presents the results of a study that explored the use of “counter-marketing” strategies in the prevention of consumer fraud.