arXiv:2511.16997v1 Announce Type: new Abstract: The emergence of AI Scientists has demonstrated remarkable potential in automating scientific research. However, current approaches largely conceptualize scientific discovery as a solitary optimization or search process, overlooking that knowledge production is inherently a social and historical endeavor. Human scientific insight stems from two distinct yet interconnected sources. First is the individual cognitive trajectory, where a researcher's unique insight is shaped by their evolving research history and stylistic preferences; another is the collective disciplinary memory, where knowledge is sedimented into vast, interconnected networks of citations and concepts. Existing LLMs still struggle to represent these structured, high-fidelity cognitive and social contexts. To bridge this gap, we introduce MirrorMind, a hierarchical cognitive architecture that integrates dual-memory representations within a three-level framework. The Individual Level constructs high-fidelity cognitive models of individual researchers by capturing their episodic, semantic, and persona memories; the Domain Level maps collective knowledge into structured disciplinary concept graphs; and the Interdisciplinary Level that acts as an orthogonal orchestration engine. Crucially, our architecture separates memory storage from agentic execution, enabling AI scientist agents to flexibly access individual memories for unique perspectives or collective structures to reason. We evaluate MirrorMind across four comprehensive tasks, including author-level cognitive simulation, complementary reasoning, cross-disciplinary collaboration promotion, and multi-agent scientific problem solving. The results show that by integrating individual cognitive depth with collective disciplinary breadth, MirrorMind moves beyond simple fact retrieval toward structural, personalized, and insight-generating scientific reasoning.

arXiv:2511.16698v1 Announce Type: cross Abstract: SNOMED CT is a biomedical ontology with a hierarchical representation of large-scale concepts. Knowledge retrieval in SNOMED CT is critical for its application, but often proves challenging due to language ambiguity, synonyms, polysemies and so on. This problem is exacerbated when the queries are out-of-vocabulary (OOV), i.e., having no equivalent matchings in the ontology. In this work, we focus on the problem of hierarchical concept retrieval from SNOMED CT with OOV queries, and propose an approach based on language model-based ontology embeddings. For evaluation, we construct OOV queries annotated against SNOMED CT concepts, testing the retrieval of the most direct subsumers and their less relevant ancestors. We find that our method outperforms the baselines including SBERT and two lexical matching methods. While evaluated against SNOMED CT, the approach is generalisable and can be extended to other ontologies. We release code, tools, and evaluation datasets at https://github.com/jonathondilworth/HR-OOV.

arXiv:2511.16846v1 Announce Type: cross Abstract: Large language models (LLMs) frequently generate responses that are lengthy and verbose, filled with redundant or unnecessary details. This diminishes clarity and user satisfaction, and it increases costs for model developers, especially with well-known proprietary models that charge based on the number of output tokens. In this paper, we introduce a novel reference-free metric for evaluating the conciseness of responses generated by LLMs. Our method quantifies non-essential content without relying on gold standard references and calculates the average of three calculations: i) a compression ratio between the original response and an LLM abstractive summary; ii) a compression ratio between the original response and an LLM extractive summary; and iii) wordremoval compression, where an LLM removes as many non-essential words as possible from the response while preserving its meaning, with the number of tokens removed indicating the conciseness score. Experimental results demonstrate that our proposed metric identifies redundancy in LLM outputs, offering a practical tool for automated evaluation of response brevity in conversational AI systems without the need for ground truth human annotations.

arXiv:2511.17432v1 Announce Type: cross Abstract: Traditional evaluation metrics for textual and visual question answering, like ROUGE, METEOR, and Exact Match (EM), focus heavily on n-gram based lexical similarity, often missing the deeper semantic understanding needed for accurate assessment. While measures like BERTScore and MoverScore leverage contextual embeddings to address this limitation, they lack flexibility in balancing sentence-level and keyword-level semantics and ignore lexical similarity, which remains important. Large Language Model (LLM) based evaluators, though powerful, come with drawbacks like high costs, bias, inconsistency, and hallucinations. To address these issues, we introduce SMILE: Semantic Metric Integrating Lexical Exactness, a novel approach that combines sentence-level semantic understanding with keyword-level semantic understanding and easy keyword matching. This composite method balances lexical precision and semantic relevance, offering a comprehensive evaluation. Extensive benchmarks across text, image, and video QA tasks show SMILE is highly correlated with human judgments and computationally lightweight, bridging the gap between lexical and semantic evaluation.

arXiv:2511.17467v1 Announce Type: cross Abstract: We propose a novel framework for persona-based language model system, motivated by the need for personalized AI agents that adapt to individual user preferences. In our approach, the agent embodies the user's "persona" (e.g. user profile or taste) and is powered by a large language model (LLM). To enable the agent to leverage rich contextual information, we introduce a Knowledge-Graph-enhanced Retrieval-Augmented Generation (Graph RAG) mechanism that constructs an LLM-derived graph index of relevant documents and summarizes communities of related information. Our framework generates personalized prompts by combining: (1) a summary of the user's historical behaviors and preferences extracted from the knowledge graph, and (2) relevant global interaction patterns identified through graph-based community detection. This dynamic prompt engineering approach allows the agent to maintain consistent persona-aligned behaviors while benefiting from collective knowledge. On the LaMP benchmark, our method improves news categorization F1 by 11.1%, movie tagging F1 by 56.1%, and reduces product rating MAE by 10.4% over prior methods. Our code is available at https://anonymous.4open.science/r/PersonaAgentwGraphRAG-DE6F

arXiv:2503.01424v4 Announce Type: replace Abstract: Research is a fundamental process driving the advancement of human civilization, yet it demands substantial time and effort from researchers. In recent years, the rapid development of artificial intelligence (AI) technologies has inspired researchers to explore how AI can accelerate and enhance research. To monitor relevant advancements, this paper presents a systematic review of the progress in this domain. Specifically, we organize the relevant studies into three main categories: hypothesis formulation, hypothesis validation, and manuscript publication. Hypothesis formulation involves knowledge synthesis and hypothesis generation. Hypothesis validation includes the verification of scientific claims, theorem proving, and experiment validation. Manuscript publication encompasses manuscript writing and the peer review process. Furthermore, we identify and discuss the current challenges faced in these areas, as well as potential future directions for research. Finally, we also offer a comprehensive overview of existing benchmarks and tools across various domains that support the integration of AI into the research process. We hope this paper serves as an introduction for beginners and fosters future research. Resources have been made publicly available at https://github.com/zkzhou126/AI-for-Research.

arXiv:2504.07139v3 Announce Type: replace Abstract: Welcome to the eighth edition of the AI Index report. The 2025 Index is our most comprehensive to date and arrives at an important moment, as AI's influence across society, the economy, and global governance continues to intensify. New in this year's report are in-depth analyses of the evolving landscape of AI hardware, novel estimates of inference costs, and new analyses of AI publication and patenting trends. We also introduce fresh data on corporate adoption of responsible AI practices, along with expanded coverage of AI's growing role in science and medicine. Since its founding in 2017 as an offshoot of the One Hundred Year Study of Artificial Intelligence, the AI Index has been committed to equipping policymakers, journalists, executives, researchers, and the public with accurate, rigorously validated, and globally sourced data. Our mission has always been to help these stakeholders make better-informed decisions about the development and deployment of AI. In a world where AI is discussed everywhere - from boardrooms to kitchen tables - this mission has never been more essential. The AI Index continues to lead in tracking and interpreting the most critical trends shaping the field - from the shifting geopolitical landscape and the rapid evolution of underlying technologies, to AI's expanding role in business, policymaking, and public life. Longitudinal tracking remains at the heart of our mission. In a domain advancing at breakneck speed, the Index provides essential context - helping us understand where AI stands today, how it got here, and where it may be headed next. Recognized globally as one of the most authoritative resources on artificial intelligence, the AI Index has been cited in major media outlets such as The New York Times, Bloomberg, and The Guardian; referenced in hundreds of academic papers; and used by policymakers and government agencies around the world.

arXiv:2409.11393v3 Announce Type: replace-cross Abstract: In an era where vast amounts of data are collected and processed from diverse sources, there is a growing demand for sophisticated AI systems capable of intelligently fusing and analyzing this information. To address these challenges, researchers have turned towards integrating tools into LLM-powered agents to enhance the overall information fusion process. However, the conjunction of these technologies and the proposed enhancements in several state-of-the-art works followed a non-unified software architecture, resulting in a lack of modularity and terminological inconsistencies among researchers. To address these issues, we propose a novel LLM-based Agent Unified Modeling Framework (LLM-Agent-UMF) that establishes a clear foundation for agent development from both functional and software architectural perspectives, developed and evaluated using the Architecture Tradeoff and Risk Analysis Framework (ATRAF). Our framework clearly distinguishes between the different components of an LLM-based agent, setting LLMs and tools apart from a new element, the core-agent, which plays the role of central coordinator. This pivotal entity comprises five modules: planning, memory, profile, action, and security -- the latter often neglected in previous works. By classifying core-agents into passive and active types based on their authoritative natures, we propose various multi-core agent architectures that combine unique characteristics of distinctive agents to tackle complex tasks more efficiently. We evaluate our framework by applying it to thirteen state-of-the-art agents, thereby demonstrating its alignment with their functionalities and clarifying overlooked architectural aspects. Moreover, we thoroughly assess five architecture variants of our framework by designing new agent architectures that combine characteristics of state-of-the-art agents to address specific goals. ...

arXiv:2511.12797v2 Announce Type: replace-cross Abstract: In-context learning (ICL) -- the capacity of a model to infer and apply abstract patterns from examples provided within its input -- has been extensively studied in large language models trained for next-token prediction on human text. In fact, prior work often attributes this emergent behavior to distinctive statistical properties in human language. This raises a fundamental question: can ICL arise organically in other sequence domains purely through large-scale predictive training? To explore this, we turn to genomic sequences, an alternative symbolic domain rich in statistical structure. Specifically, we study the Evo2 genomic model, trained predominantly on next-nucleotide (A/T/C/G) prediction, at a scale comparable to mid-sized LLMs. We develop a controlled experimental framework comprising symbolic reasoning tasks instantiated in both linguistic and genomic forms, enabling direct comparison of ICL across genomic and linguistic models. Our results show that genomic models, like their linguistic counterparts, exhibit log-linear gains in pattern induction as the number of in-context demonstrations increases. To the best of our knowledge, this is the first evidence of organically emergent ICL in genomic sequences, supporting the hypothesis that ICL arises as a consequence of large-scale predictive modeling over rich data. These findings extend emergent meta-learning beyond language, pointing toward a unified, modality-agnostic view of in-context learning.

npj Digital Medicine, Published online: 24 November 2025; doi:10.1038/s41746-025-02162-4

On the public dissemination and open sourcing of ultrasound resources, datasets and deep learning models

npj Digital Medicine, Published online: 24 November 2025; doi:10.1038/s41746-025-02095-y

Multimodal analysis of whole slide images in colorectal cancer

Nature, Published online: 24 November 2025; doi:10.1038/d41586-025-03857-0

The advent of artificial intelligence might be just the latest stage in a guiding biological process that has produced ever more complex, mutually dependent organisms over the history of life.

Most AI benchmarks measure intelligence and instruction-following rather than psychological safety. Humane Bench evaluates models based on core principles of human flourishing, prioritizing well-being, and respecting user attention.

In 2017, fresh off a PhD on theoretical chemistry, John Jumper heard rumors that Google DeepMind had moved on from building AI that played games with superhuman skill and was starting up a secret project to predict the structures of proteins. He applied for a job.

Just three years later, Jumper celebrated a stunning win that few had seen coming. With CEO Demis Hassabis, he had co-led the development of an AI system called AlphaFold 2 that was able to predict the structures of proteins to within the width of an atom, matching the accuracy of painstaking techniques used in the lab, and doing it many times faster—returning results in hours instead of months.

AlphaFold 2 had cracked a 50-year-old grand challenge in biology. “This is the reason I started DeepMind,” Hassabis told me a few years ago. “In fact, it’s why I’ve worked my whole career in AI.” In 2024, Jumper and Hassabis shared a Nobel Prize in chemistry.

It was five years ago this week that AlphaFold 2’s debut took scientists by surprise. Now that the hype has died down, what impact has AlphaFold really had? How are scientists using it? And what’s next? I talked to Jumper (as well as a few other scientists) to find out.

“It’s been an extraordinary five years,” Jumper says, laughing: “It’s hard to remember a time before I knew tremendous numbers of journalists.”

AlphaFold 2 was followed by AlphaFold Multimer, which could predict structures that contained more than one protein, and then AlphaFold 3, the fastest version yet. Google DeepMind also let AlphaFold loose on UniProt, a vast protein database used and updated by millions of researchers around the world. It has now predicted the structures of some 200 million proteins, almost all that are known to science.

Despite his success, Jumper remains modest about AlphaFold’s achievements. “That doesn’t mean that we’re certain of everything in there,” he says. “It’s a database of predictions, and it comes with all the caveats of predictions.”

A hard problem

Proteins are the biological machines that make living things work. They form muscles, horns, and feathers; they carry oxygen around the body and ferry messages between cells; they fire neurons, digest food, power the immune system; and so much more. But understanding exactly what a protein does (and what role it might play in various diseases or treatments) involves figuring out its structure—and that’s hard.

Proteins are made from strings of amino acids that chemical forces twist up into complex knots. An untwisted string gives few clues about the structure it will form. In theory, most proteins could take on an astronomical number of possible shapes. The task is to predict the correct one.

Jumper and his team built AlphaFold 2 using a type of neural network called a transformer, the same technology that underpins large language models. Transformers are very good at paying attention to specific parts of a larger puzzle.

But Jumper puts a lot of the success down to making a prototype model that they could test quickly. “We got a system that would give wrong answers at incredible speed,” he says. “That made it easy to start becoming very adventurous with the ideas you try.”

They stuffed the neural network with as much information about protein structures as they could, such as how proteins across certain species have evolved similar shapes. And it worked even better than they expected. “We were sure we had made a breakthrough,” says Jumper. “We were sure that this was an incredible advance in ideas.”

What he hadn’t foreseen was that researchers would download his software and start using it straight away for so many different things. Normally, it’s the thing a few iterations down the line that has the real impact, once the kinks have been ironed out, he says: “I’ve been shocked at how responsibly scientists have used it, in terms of interpreting it, and using it in practice about as much as it should be trusted in my view, neither too much nor too little.”

Any projects stand out in particular? 

Honeybee science

Jumper brings up a research group that uses AlphaFold to study disease resistance in honeybees. “They wanted to understand this particular protein as they look at things like colony collapse,” he says. “I never would have said, ‘You know, of course AlphaFold will be used for honeybee science.’”

He also highlights a few examples of what he calls off-label uses of AlphaFold—“in the sense that it wasn’t guaranteed to work”—where the ability to predict protein structures has opened up new research techniques. “The first is very obviously the advances in protein design,” he says. “David Baker and others have absolutely run with this technology.”

Baker, a computational biologist at the University of Washington, was a co-winner of last year’s chemistry Nobel, alongside Jumper and Hassabis, for his work on creating synthetic proteins to perform specific tasks—such as treating disease or breaking down plastics—better than natural proteins can.

Baker and his colleagues have developed their own tool based on AlphaFold, called RoseTTAFold. But they have also experimented with AlphaFold Multimer to predict which of their designs for potential synthetic proteins will work.    

“Basically, if AlphaFold confidently agrees with the structure you were trying to design then you make it and if AlphaFold says ‘I don’t know,’ you don’t make it. That alone was an enormous improvement.” It can make the design process 10 times faster, says Jumper.

Another off-label use that Jumper highlights: Turning AlphaFold into a kind of search engine. He mentions two separate research groups that were trying to understand exactly how human sperm cells hooked up with eggs during fertilization. They knew one of the proteins involved but not the other, he says: “And so they took a known egg protein and ran all 2,000 human sperm surface proteins, and they found one that AlphaFold was very sure stuck against the egg.” They were then able to confirm this in the lab.

“This notion that you can use AlphaFold to do something you couldn’t do before—you would never do 2,000 structures looking for one answer,” he says. “This kind of thing I think is really extraordinary.”

Five years on

When AlphaFold 2 came out, I asked a handful of early adopters what they made of it. Reviews were good, but the technology was too new to know for sure what long-term impact it might have. I caught up with one of those people to hear his thoughts five years on.

Kliment Verba is a molecular biologist who runs a lab at the University of California, San Francisco. “It’s an incredibly useful technology, there’s no question about it,” he tells me. “We use it every day, all the time.”

But it’s far from perfect. A lot of scientists use AlphaFold to study pathogens or to develop drugs. This involves looking at interactions between multiple proteins or between proteins and even smaller molecules in the body. But AlphaFold is known to be less accurate at making predictions about multiple proteins or their interaction over time.

Verba says he and his colleagues have been using AlphaFold long enough to get used to its limitations. “There are many cases where you get a prediction and you have to kind of scratch your head,” he says. “Is this real or is this not? It’s not entirely clear—it’s sort of borderline.”

“It’s sort of the same thing as ChatGPT,” he adds. “You know—it will bullshit you with the same confidence as it would give a true answer.”

Still, Verba’s team uses AlphaFold (both 2 and 3, because they have different strengths, he says) to run virtual versions of their experiments before running them in the lab. Using AlphaFold’s results, they can narrow down the focus of an experiment—or decide that it’s not worth doing.

It can really save time, he says: “It hasn’t really replaced any experiments, but it’s augmented them quite a bit.”

New wave  

AlphaFold was designed to be used for a range of purposes. Now multiple startups and university labs are building on its success to develop a new wave of tools more tailored to drug discovery. This year, a collaboration between MIT researchers and the AI drug company Recursion produced a model called Boltz-2, which predicts not only the structure of proteins but also how well potential drug molecules will bind to their target.  

Last month, the startup Genesis Molecular AI released another structure prediction model called Pearl, which the firm claims is more accurate than AlphaFold 3 for certain queries that are important for drug development. Pearl is interactive, so that drug developers can feed any additional data they may have to the model to guide its predictions.

AlphaFold was a major leap, but there’s more to do, says Evan Feinberg, Genesis Molecular AI’s CEO: “We’re still fundamentally innovating, just with a better starting point than before.”

Genesis Molecular AI is pushing margins of error down from less than two angstroms, the de facto industry standard set by AlphaFold, to less than one angstrom—one 10-millionth of a millimeter, or the width of a single hydrogen atom.

“Small errors can be catastrophic for predicting how well a drug will actually bind to its target,” says Michael LeVine, vice president of modeling and simulation at the firm. That’s because chemical forces that interact at one angstrom can stop doing so at two. “It can go from ‘They will never interact’ to ‘They will,’” he says.

With so much activity in this space, how soon should we expect new types of drugs to hit the market? Jumper is pragmatic. Protein structure prediction is just one step of many, he says: “This was not the only problem in biology. It’s not like we were one protein structure away from curing any diseases.”

Think of it this way, he says. Finding a protein’s structure might previously have cost $100,000 in the lab: “If we were only a hundred thousand dollars away from doing a thing, it would already be done.”

At the same time, researchers are looking for ways to do as much as they can with this technology, says Jumper: “We’re trying to figure out how to make structure prediction an even bigger part of the problem, because we have a nice big hammer to hit it with.”

In other words, make everything into nails? “Yeah, let’s make things into nails,” he says. “How do we make this thing that we made a million times faster a bigger part of our process?”

What’s next?

Jumper’s next act? He wants to fuse the deep but narrow power of AlphaFold with the broad sweep of LLMs.  

“We have machines that can read science. They can do some scientific reasoning,” he says. “And we can build amazing, superhuman systems for protein structure prediction. How do you get these two technologies to work together?”

That makes me think of a system called AlphaEvolve, which is being built by another team at Google DeepMind. AlphaEvolve uses an LLM to generate possible solutions to a problem and a second model to check them, filtering out the trash. Researchers have already used AlphaEvolve to make a handful of practical discoveries in math and computer science.    

Is that what Jumper has in mind? “I won’t say too much on methods, but I’ll be shocked if we don’t see more and more LLM impact on science,” he says. “I think that’s the exciting open question that I’ll say almost nothing about. This is all speculation, of course.”

Jumper was 39 when he won his Nobel Prize. What’s next for him?

“It worries me,” he says. “I believe I’m the youngest chemistry laureate in 75 years.” 

He adds: “I’m at the midpoint of my career, roughly. I guess my approach to this is to try to do smaller things, little ideas that you keep pulling on. The next thing I announce doesn’t have to be, you know, my second shot at a Nobel. I think that’s the trap.”

Oncogene, Published online: 24 November 2025; doi:10.1038/s41388-025-03648-x

Ten years of oncogene editorship: a decade of transformation

Background: The rapid advancement of digital health technologies (DHTs) offers substantial potential for improving healthcare access, yet it simultaneously risks exacerbating existing inequities for marginalized populations. Previous research on the digital divide has often treated individuals with blindness as a homogenous group, primarily focusing on barriers related to digital access and skills. However, less is known about the nuanced experiences of specific subgroups, such as educated and digitally literate young adults. This study focuses on this demographic to understand how their advanced digital capabilities interact with systemic and infrastructural barriers in healthcare. Objective: This qualitative study aimed to explore the lived healthcare experiences of educated young adults with blindness in China, specifically identifying how DHTs simultaneously contribute to their empowerment and exclusion. Methods: Eligible participants were educated young adults with blindness in China (aged 18-30 years, Mandarin speakers, smartphone users, and holding or pursuing higher education). A total of 12 semi-structured interviews were conducted in Mandarin during September 2024. All interviews were audio-recorded and transcribed verbatim. An inductive thematic analysis was employed to interpret the data and identify key themes. Results: Participants’ experiences highlighted an “empowered but excluded” dynamic. Seven key themes emerged, categorized into empowerment and exclusion. Empowerment themes included: (1) digital platforms empowering self-management and healthcare access, where DHTs enabled independent appointment booking and access to comprehensive health information; and (2) digital platforms empowering for finding medical visit companions, facilitating the discovery of companions for physical and emotional support. Exclusion themes comprised: (3) inaccessible online appointment systems, due to non-inclusive designs; (4) inaccessible healthcare environments and information formats, stemming from non-accessible self-service machines and written materials; (5) lack of provider competencies in respecting patient autonomy, as providers often assumed digital incompetence; (6) data privacy and security concerns, heightened by increased digitalization and reliance on assistive tools; and (7) challenges related to the quality and consistency of online companion support, highlighting the limitations of platform-based assistance. Conclusions: Our findings reveal an “empowered but excluded” dynamic: the potential for digital empowerment and enhanced independence is often curtailed by systematic barriers. Addressing this necessitates a multifaceted approach: enhancing technological accessibility through robust standards adherence and inclusive co-design processes; improving healthcare provider competencies in patient-centered care via targeted training; and empowering educated young blind adults by building their capacity for self-determination to achieve equitable healthcare access.

Background: Digital interventions can provide cost-effective, quality health care for patients with chronic conditions. Patients with chronic conditions often are burdened by a substantial load of adhering to a treatment regimen and suffer from impacts on their function and well-being. This treatment burden has consequences for treatment adherence and disease outcomes. Digital interventions have the potential to alleviate the burden, but they also may cause new challenges and an increased workload for the patient. Previous reviews have examined digital interventions or treatment burden separately, but there is a lack of systematic reviews on the intersection of digital interventions, treatment burden, and chronic conditions. Objective: This systematic review aimed to evaluate the evidence of how digital interventions impact the treatment burden experienced by people with chronic conditions, and to assess the quality of this evidence. Methods: We searched databases PubMed, Scopus, Web of Science, ACM, PubMed Central, and CINAHL for articles published between January 1, 2013, and June 17, 2025. We included studies that had key topics related to chronic conditions, treatment burden, and digital interventions. A total of 2 reviewers independently screened the articles in 2 stages, extracted data on study design, participant characteristics, intervention type, and treatment burden outcomes from included articles, and assessed their quality using the Critical Appraisal tools from the Joanna Briggs Institute. A convergent integrated approach was used for data synthesis and integration, where quantitative data were converted into qualitative data, and the qualitative and quantitative evidence were analyzed and categorized together. Results: We included 46 relevant studies in total. We categorized the interventions into 4 types: Telehealth, informational resources, self-management tools, and facilitated tools. The results of this study indicate that digital interventions mostly support patients with chronic conditions with their treatment burden, with minor concerns of increasing treatment burden. The main benefits are support with self-management, informational support, and easier ways to contact health care professionals. The main concerns were accessibility issues, time-consuming tools, and causing fear and anxiety. Conclusions: Our findings demonstrate how treatment burden is a relevant concept for future digital health care research and practice. Digital interventions can help patients with their treatment burden by supporting self-management, improving access to health care, improving patients’ experience, and addressing relevant concerns. More research is needed about conditions with low or medium initial treatment burden.

Background: The application of large language models (LLMs) in health care holds significant potential for enhancing patient care and advancing medical research. However, the protection of patient privacy remains a critical issue, especially when handling patient health information (PHI). Objective: This scoping review aims to evaluate the adequacy of current approaches and identify areas in need of improvement to ensure robust patient privacy protection in the existing studies about PHI-LLMs within the health care domain. Methods: A search of the literature published from January 1, 2022, to July 20, 2025, was performed on July 20, 2025, using 2 databases (PubMed and Embase). This scoping review focused on the following three research questions: (1) What studies on the development and application of LLMs using PHI currently exist within the health care domain? (2) What patient privacy considerations are addressed in existing PHI-LLMs research, and are these measures sufficient? (3) How can future research on the development and application of LLMs using PHI better protect patient privacy? Studies were included if they focused on the development and application of LLMs within health care using PHI, encompassing activities such as model construction, fine-tuning, optimization, testing, and performance comparison. Eligible literature comprised original research articles written in English. Conversely, studies were excluded if they used publicly available datasets, under the assumption that such data have been adequately deidentified. Additionally, non-English publications, reviews, abstracts, incomplete reports, and preprints were excluded from the review due to the lack of rigorous peer review. Results: This study systematically identified 9823 studies on PHI-LLM and included 464 studies published between 2022 and 2025. Among the 464 studies, (1) a small number of studies neglected ethical review (n=45, 9.7%) and patient informed consent (n=148, 31.9%) during the research process, (2) more than a third of the studies (n=178, 38.4%) failed to report whether to implement effective measures to protect PHI, and (3) there was a significant lack of transparency and comprehensive detail in anonymization and deidentification methods. Conclusions: We propose comprehensive recommendations across 3 phases—study design, implementation, and reporting—to strengthen patient privacy protection and transparency in PHI-LLM. This study emphasizes the urgent need for the development of stricter regulatory frameworks and the adoption of advanced privacy protection technologies to effectively safeguard PHI. It is anticipated that future applications of LLMs in the health care field will achieve a balance between innovation and robust patient privacy protection, thereby enhancing ethical standards and scientific credibility.