AI Health Roundup – July 17, 2026

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.

July 17, 2026

In this week’s Duke AI Health Friday Roundup: Anthropic researchers peer under the hood of Claude’s “J-Space”; new cholesterol therapy approved by FDA; probing the dynamics of prestige and peer review; the state of play for digital twins in medicine; CMS clamps down on remote monitoring; Cyclospora continues its rampage across the US; AI, scientific publishing, and the return of Taylorism; Meta employees allege firing decisions were made by AI; much more:

AI, STATISTICS & DATA SCIENCE

In a selective focus photograph, a hand holds a camera lens between thumb and forefinger. The blurry background of a shoreline scene with docks and pilings is brought into focus inside the lens, which also imparts a pinkish or purplish tint to the scene. Image credit: Stephen/Unsplash
Image credit: Stephen/Unsplash
  • “Anthropic’s J-lens works in a similar way but picks out words that an LLM is likely to say at some point in the near future, not necessarily straight away. What that reveals in practice are words that are related to the response an LLM is working on but that might not actually end up being part of that response by the time the math in the middle layers has run its course….if Claude were a person (it’s not), you might say that the J-lens gives clues about what it is thinking about at different levels of the book stack but not saying out loud.” At MIT Technology Review, Will Douglas Heaven reports on a new paper from Anthropic describing strange doings afoot in “J-Space” in the Claude transformer model.
  • “…this field is in a rapid development phase driven by artificial intelligence (AI) and interdisciplinary integration, demonstrating significant potential in enhancing diagnostic accuracy, efficiency, and driving paradigm shifts. However, the comprehensive analysis also highlights the systemic challenges in transitioning from technological prototypes to large-scale clinical translation.” A scoping review and analysis published in the journal Artificial Intelligence in Medicine by Shen and colleagues provides a snapshot at the current state of “digital twins” in medicine.
  • “…scientific publishing is starting to experience some Taylorism with the insertion of AI. Rigorous human checking of AI-generated research papers is creating bottlenecks as publishers strive to maintain the integrity of the scientific record. The challenge is requiring even more human effort, making the whole endeavor slower and more expensive.” Science editor in chief Holden Thorpe confronts the specter of Taylorism as AI becomes a force in scientific scholarly publishing.
  • “…the findings got limited coverage in the press, and the story on bioengineer.org is basically an AI-generated copy of the release by the Endocrine Society….Which, though it’s been shared a measly (supposed) 77 times, is probably not all that damaging to science communication, right? I can’t imagine many people frequent bioengineer.org and if they do, and they see the molle, methds and meda graph above, hopefully they will just click away…What is concerning, and damaging, is the way bioengineer has replaced Actual Real Human Scientists with ghosts.” At No Breakthroughs, Jackson Ryan documents the further encroachments of AI slop on science news sites.
  • “Meta’s AI-fueled layoffs of 8,000 employees targeted workers with disabilities and those who took protected medical or family leaves, alleged a lawsuit filed by 26 employees who were selected for termination. Meta used internal AI tools to select employees for layoffs, according to the complaint filed yesterday by 26 “Doe” plaintiffs in US District Court for the Northern District of California.” Ars Technica’s Jon Brodkin has the story of a lawsuit filed against Meta alleging the use of AIs to make sweeping decisions about layoffs.

BASIC SCIENCE, CLINICAL RESEARCH & PUBLIC HEALTH

Green lettuce leaves on a white background. Image credit: Petr/Unsplash
Image credit: Petr/Unsplash
  • “Since 1 May, 1,645 infections with the parasite Cyclospora cayetanensis have been confirmed, and more than 5,100 further cases are under investigation. These figures — which have accumulated in just a few weeks — have quickly surpassed the total number of cases typically seen in a year, says Joel Barratt, a molecular parasitologist at Emory School of Medicine in Atlanta, Georgia. More than 100 people with C. cayetanensis infections have been hospitalized, but no deaths have been reported.” Nature’s Max Kozlov covers the current state of the US cyclosporiasis outbreak, which is being complicated reductions in infection-tracking staff at federal agencies.
  • “Left-handedness does not seem to be purely genetic. Two left-handed parents have a left-handed child only 25% to 30% of the time. If one identical twin is left-handed, there’s only a 20% to 30% chance the other is too. This suggests a genetic component alongside some developmental randomness….Also, the rightward bias of human handedness is unique. Some other mammals can be right-footed or left-pawed, but there’s no statistical imbalance toward left or right across the species, except for our extreme 10-90 split.” An article by Quanta’s Natalie Wolchover examines the surprisingly difficult question of why some people are left-handed.
  • “The Food and Drug Administration approved a daily pill on Thursday that can lower cholesterol levels far below what can be achieved with statins, the cheap cholesterol-reducing pills….The drug, enlicitide, whose brand name is Lipfendra, is made by the pharmaceutical company Merck. Clinical trials have shown that it can bring levels of LDL — the dangerous type of cholesterol — down to 50 or 60 or even lower.” The New York Times’ Gina Kolata reports that the FDA has approved an oral PCSK9 inhibitor drug for lowering cholesterol, one that will be cheaper than existing injectable PCSK9 inhibitor options.

COMMUNICATIONS & Policy

This image shows a large white, traditional, old building. The top half of the building represents the humanities (which is symbolised by the embedded text from classic literature which is faintly shown ontop the building). The bottom section of the building is embossed with mathematical formulas to represent the sciences. The middle layer of the image is heavily pixelated. On the steps at the front of the building there is a group of scholars, wearing formal suits and tie attire, who are standing around at the enternace talking and some of them are sitting on the steps. There are two stone, statute-like hands that are stretching the building apart from the left side. In the forefront of the image, there are 8 students - which can only be seen from the back. Their graduation gowns have bright blue hoods and they all look as though they are walking towards the old building which is in the background at a distance. There are a mix of students in the foreground. Image credit: Zoya Yasmine / https://betterimagesofai.org / https://creativecommons.org/licenses/by/4.0/
Image credit: Zoya Yasmine / https://betterimagesofai.org / https://creativecommons.org/licenses/by/4.0/
  • “…we find strong selective effects associated with higher institutional prestige, larger team size, and certain topics and countries. Corresponding authors who are men exhibit a small but significant advantage at Science, while authors based in China have a significant disadvantage. These associations are generally stronger in editorial review than in peer review, even as final editorial decisions correlate strongly with reviewer advice.” A research article published in Science Advances by Zhang and colleagues offers an analysis of the inner workings of peer review and editorial decision-making at two high-impact scientific journals.
  • “Medicare has covered remote patient monitoring since 2018, and payments ballooned to over $500 million in 2024. The proposed change follows widespread concern from the health department’s watchdog, academics, and insurers that the current remote monitoring system is paying for low-value services. If finalized, the rule would impact a large percentage of remote monitoring care as it’s delivered today.” STAT News’ Mario Aguilar reports that the Centers for Medicare and Medicaid Services is clamping down on remote patient monitoring services provided by third-party vendors.
  • Duke doctoral student Max Bucklan, part of the genomics-focused Gregory Lab, is among the winners of the Lasker Foundation’s annual contest for videos explaining the work of early-career scientists.
  • “The analysis, done by researchers at the University of North Carolina at Chapel Hill in partnership with STAT, used a large language model to classify whether a comment was supportive or in opposition and to identify themes mentioned in comments. It found that about 95% were in opposition and just 1% supported the proposed changes to the ‘Uniform Guidance.’ A total of 496,769 public comments were submitted before the deadline of Monday at midnight, and an analysis of the 52,322 comments that have been posted in full show an overwhelming rebuke of the proposal.” STAT News’ Anil Oza and J. Emory Parker report on early tabulations of comments registered as part of the public comment period for a new federal rule by the Office of Management and Budget that would place funding decisions a large swath of federal grant funding, including most of the current research funding system, under the control of political appointees.

Sites@Duke Express is powered by WordPress. Read the Sites@Duke Express policies and FAQs, or request help.