arXiv:2601.06223v1 Announce Type: cross Abstract: This paper presents a conceptual and operational framework for developing and operating safe and trustworthy AI agents based on a Three-Pillar Model grounded in transparency, accountability, and trustworthiness. Building on prior work in Human-in-the-Loop systems, reinforcement learning, and collaborative AI, the framework defines an evolutionary path toward autonomous agents that balances increasing automation with appropriate human oversight. The paper argues that safe agent autonomy must be achieved through progressive validation, analogous to the staged development of autonomous driving, rather than through immediate full automation. Transparency and accountability are identified as foundational requirements for establishing user trust and for mitigating known risks in generative AI systems, including hallucinations, data bias, and goal misalignment, such as the inversion problem. The paper further describes three ongoing work streams supporting this framework: public deliberation on AI agents conducted by the Stanford Deliberative Democracy Lab, cross-industry collaboration through the Safe AI Agent Consortium, and the development of open tooling for an agent operating environment aligned with the Three-Pillar Model. Together, these contributions provide both conceptual clarity and practical guidance for enabling the responsible evolution of AI agents that operate transparently, remain aligned with human values, and sustain societal trust.

arXiv:2511.12439v2 Announce Type: replace Abstract: Online health resources and large language models (LLMs) are increasingly used as a first point of contact for medical decision-making, yet their reliability in healthcare remains limited by low accuracy, lack of transparency, and susceptibility to unverified information. We introduce a proof-of-concept conversational self-triage system that guides LLMs with 100 clinically validated flowcharts from the American Medical Association, providing a structured and auditable framework for patient decision support. The system leverages a multi-agent framework consisting of a retrieval agent, a decision agent, and a chat agent to identify the most relevant flowchart, interpret patient responses, and deliver personalized, patient-friendly recommendations, respectively. Performance was evaluated at scale using synthetic datasets of simulated conversations. The system achieved 95.29% top-3 accuracy in flowchart retrieval (N=2,000) and 99.10% accuracy in flowchart navigation across varied conversational styles and conditions (N=37,200). By combining the flexibility of free-text interaction with the rigor of standardized clinical protocols, this approach demonstrates the feasibility of transparent, accurate, and generalizable AI-assisted self-triage, with potential to support informed patient decision-making while improving healthcare resource utilization.

arXiv:2512.05765v1 Announce Type: new Abstract: Influential critiques argue that Large Language Models (LLMs) are a dead end for AGI: "mere pattern matchers" structurally incapable of reasoning or planning. We argue this conclusion misidentifies the bottleneck: it confuses the ocean with the net. Pattern repositories are the necessary System-1 substrate; the missing component is a System-2 coordination layer that selects, constrains, and binds these patterns. We formalize this layer via UCCT, a theory of semantic anchoring that models reasoning as a phase transition governed by effective support (rho_d), representational mismatch (d_r), and an adaptive anchoring budget (gamma log k). Under this lens, ungrounded generation is simply an unbaited retrieval of the substrate's maximum likelihood prior, while "reasoning" emerges when anchors shift the posterior toward goal-directed constraints. We translate UCCT into architecture with MACI, a coordination stack that implements baiting (behavior-modulated debate), filtering (Socratic judging), and persistence (transactional memory). By reframing common objections as testable coordination failures, we argue that the path to AGI runs through LLMs, not around them.

Background: Sleep is an important component of human health and can be measured longitudinally using digital activity trackers. Further, decentralized digital research has the potential to provide a real-world picture of sleep in large populations. Objective: This study examined whether longitudinal sleep patterns from activity trackers could predict risk of obstructive sleep apnea (OSA) and hypertension, as defined the Berlin questionnaire and self report, respectively. Methods: We recruited adults ≥18 years nationwide to join our sleep-focused smartphone-based study, called the Research Framework for Exploring Sleep Health (REFRESH). Our sample of N= 391 adults is comprised predominately of females (68%; n = 247 out of 364) at a mean age of 48 years (standard deviation (SD) = 13.62). Participants were asked to fill out health-related surveys, including the Berlin questionnaire, and the Horne-Ostberg questionnaire for chronotype. Participants were asked to link their own activity tracker to the application to collect longitudinal sleep data. Results: We analyzed sleep data from 391 participants, among a predominately White (65%; n = 231 out of 353) followed by multiracial (17%; n = 61 out of 353) and Hispanic or Latino (6.5%; n = 23 out of 353) cohort. Collinearity testing showed that OSA risk and self-reported hypertension could be considered independently. Holding body mass index (BMI) at fixed value, the odds of having high OSA risk increased by 159% for every one-hour increase in weekday sleep variability (odds ratio (OR) = 2.592, 95% confidence interval (CI) [1.613, 4.400]; P

arXiv:2506.02139v5 Announce Type: replace Abstract: We propose semantic anchoring, a unified account of how large language models turn pretrained capacity into goal-directed behavior: external structure (in-context examples, retrieval, or light tuning) binds the model's latent patterns to desired targets. Unified Contextual Control Theory (UCCT) formalizes this via anchoring strength $S = \rho_d - d_r - \log k$, where $\rho_d$ measures target cohesion in representation space, $d_r$ measures mismatch from prior knowledge, and $k$ is the anchor budget. UCCT predicts threshold-like performance flips and strictly generalizes in-context learning, reading retrieval and fine-tuning as anchoring variants. Three controlled studies provide evidence. Experiment 1 demonstrates cross-domain anchoring rebinding strong priors in text and vision. Experiment 2 varies representational familiarity via numeral bases (base-10/8/9) at fixed complexity, yielding ordered thresholds and transfer patterns tracking $\rho_d$, $d_r$, and $S$. Experiment 3 establishes a geometry-to-behavior correlate: layer-wise peak anchoring and trajectory area predict few-shot thresholds $\theta_{50}$. UCCT offers testable theory and practical metrics for optimizing prompts, retrieval, and tuning.

arXiv:2511.18274v1 Announce Type: cross Abstract: Digital health interventions are increasingly used in physical and occupational therapy to deliver home exercise programs via sensor equipped devices such as smartphones, enabling remote monitoring of adherence and performance. However, digital interventions are typically programmed as software before clinical encounters as libraries of parametrized exercise modules targeting broad patient populations. At the point of care, clinicians can only select modules and adjust a narrow set of parameters like repetitions, so patient specific needs that emerge during encounters, such as distinct movement limitations, and home environments, are rarely reflected in the software. We evaluated a digital intervention paradigm that uses large language models (LLMs) to translate clinicians' exercise prescriptions into intervention software. In a prospective single arm feasibility study with 20 licensed physical and occupational therapists and a standardized patient, clinicians created 40 individualized upper extremity programs (398 instructions) that were automatically translated into executable software. Our results show a 45% increase in the proportion of personalized prescriptions that can be implemented as software compared with a template based benchmark, with unanimous consensus among therapists on ease of use. The LLM generated software correctly delivered 99.78% (397/398) of instructions as prescribed and monitored performance with 88.4% (352/398) accuracy, with 90% (18/20) of therapists judged it safe to interact with patients, and 75% (15/20) expressed willingness to adopt it. To our knowledge, this is the first prospective evaluation of clinician directed intervention software generation with LLMs in healthcare, demonstrating feasibility and motivating larger trials to assess clinical effectiveness and safety in real patient populations.

arXiv:2506.09095v4 Announce Type: replace-cross Abstract: Foundation models (FMs) are changing the way medical images are analyzed by learning from large collections of unlabeled data. Instead of relying on manually annotated examples, FMs are pre-trained to learn general-purpose visual features that can later be adapted to specific clinical tasks with little additional supervision. In this review, we examine how FMs are being developed and applied in pathology, radiology, and ophthalmology, drawing on evidence from over 150 studies. We explain the core components of FM pipelines, including model architectures, self-supervised learning methods, and strategies for downstream adaptation. We also review how FMs are being used in each imaging domain and compare design choices across applications. Finally, we discuss key challenges and open questions to guide future research.

arXiv:2511.12439v1 Announce Type: new Abstract: Online health resources and large language models (LLMs) are increasingly used as a first point of contact for medical decision-making, yet their reliability in healthcare remains limited by low accuracy, lack of transparency, and susceptibility to unverified information. We introduce a proof-of-concept conversational self-triage system that guides LLMs with 100 clinically validated flowcharts from the American Medical Association, providing a structured and auditable framework for patient decision support. The system leverages a multi-agent framework consisting of a retrieval agent, a decision agent, and a chat agent to identify the most relevant flowchart, interpret patient responses, and deliver personalized, patient-friendly recommendations, respectively. Performance was evaluated at scale using synthetic datasets of simulated conversations. The system achieved 95.29% top-3 accuracy in flowchart retrieval (N=2,000) and 99.10% accuracy in flowchart navigation across varied conversational styles and conditions (N=37,200). By combining the flexibility of free-text interaction with the rigor of standardized clinical protocols, this approach demonstrates the feasibility of transparent, accurate, and generalizable AI-assisted self-triage, with potential to support informed patient decision-making while improving healthcare resource utilization.

arXiv:2510.23639v2 Announce Type: replace-cross Abstract: This paper introduces an innovative Electronic Health Record (EHR) foundation model that integrates Polygenic Risk Scores (PRS) as a foundational data modality, moving beyond traditional EHR-only approaches to build more holistic health profiles. Leveraging the extensive and diverse data from the All of Us (AoU) Research Program, this multimodal framework aims to learn complex relationships between clinical data and genetic predispositions. The methodology extends advancements in generative AI to the EHR foundation model space, enhancing predictive capabilities and interpretability. Evaluation on AoU data demonstrates the model's predictive value for the onset of various conditions, particularly Type 2 Diabetes (T2D), and illustrates the interplay between PRS and EHR data. The work also explores transfer learning for custom classification tasks, showcasing the architecture's versatility and efficiency. This approach is pivotal for unlocking new insights into disease prediction, proactive health management, risk stratification, and personalized treatment strategies, laying the groundwork for more personalized, equitable, and actionable real-world evidence generation in healthcare.

arXiv:2510.23639v1 Announce Type: cross Abstract: This paper introduces an innovative Electronic Health Record (EHR) foundation model that integrates Polygenic Risk Scores (PRS) as a foundational data modality, moving beyond traditional EHR-only approaches to build more holistic health profiles. Leveraging the extensive and diverse data from the All of Us (AoU) Research Program, this multimodal framework aims to learn complex relationships between clinical data and genetic predispositions. The methodology extends advancements in generative AI to the EHR foundation model space, enhancing predictive capabilities and interpretability. Evaluation on AoU data demonstrates the model's predictive value for the onset of various conditions, particularly Type 2 Diabetes (T2D), and illustrates the interplay between PRS and EHR data. The work also explores transfer learning for custom classification tasks, showcasing the architecture's versatility and efficiency. This approach is pivotal for unlocking new insights into disease prediction, proactive health management, risk stratification, and personalized treatment strategies, laying the groundwork for more personalized, equitable, and actionable real-world evidence generation in healthcare.

arXiv:2505.18087v2 Announce Type: replace-cross Abstract: Recent progress in Large Vision-Language Models (LVLMs) has enabled promising applications in medical tasks, such as report generation and visual question answering. However, existing benchmarks focus mainly on the final diagnostic answer, offering limited insight into whether models engage in clinically meaningful reasoning. To address this, we present CheXStruct and CXReasonBench, a structured pipeline and benchmark built on the publicly available MIMIC-CXR-JPG dataset. CheXStruct automatically derives a sequence of intermediate reasoning steps directly from chest X-rays, such as segmenting anatomical regions, deriving anatomical landmarks and diagnostic measurements, computing diagnostic indices, and applying clinical thresholds. CXReasonBench leverages this pipeline to evaluate whether models can perform clinically valid reasoning steps and to what extent they can learn from structured guidance, enabling fine-grained and transparent assessment of diagnostic reasoning. The benchmark comprises 18,988 QA pairs across 12 diagnostic tasks and 1,200 cases, each paired with up to 4 visual inputs, and supports multi-path, multi-stage evaluation including visual grounding via anatomical region selection and diagnostic measurements. Even the strongest of 12 evaluated LVLMs struggle with structured reasoning and generalization, often failing to link abstract knowledge with anatomically grounded visual interpretation. The code is available at https://github.com/ttumyche/CXReasonBench

Nature, Published online: 20 October 2025; doi:10.1038/s41586-025-09713-5

Parity and lactation induce T cell mediated breast cancer protection

Nature Medicine, Published online: 30 September 2025; doi:10.1038/s41591-025-03933-y

In a prospective study enrolling 1,222 patients from 22 emergency departments, a device using a machine-learning-based signature of blood mRNAs demonstrated clinically acceptable performance to diagnose bacterial and viral infections and to predict the all-cause need for critical care interventions within 7 days, with benchmark to established biomarkers and risk scores.

Nature, Published online: 08 January 2025; doi:10.1038/s41586-024-08388-8

Results from a comprehensive evaluation of the function of BRCA2 variants, particularly variants of uncertain significance, provide a useful resource to improve the clinical management of individuals who carry such genetic variants.

Nature, Published online: 08 May 2024; doi:10.1038/s41586-024-07370-8

Enhanced light–molecule interactions in high-finesse fibre-based Fabry–Pérot microcavities are used to detect and profile individual unlabelled solution-phase biomolecules, leading to potential applications in the life and chemical sciences.

Characterization of microbiome genomes at the species level in over 4,000 metastatic tumor biopsies identifies the distribution and diversity features of tumor-resident bacterial DNA at a pan-cancer scale, highlighting the associations between microbial community dynamics and tumor immunity and immunotherapy efficacy.

Nature, Published online: 08 April 2024; doi:10.1038/s41586-024-07379-z

Tumor-selective activity of RAS-GTP inhibition in pancreatic cancer

Immunotherapy holds strong promise for cancer treatment but at present benefits only a small proportion of cases. A pan-cancer analysis of the immune landscape in more than 1,000 tumors across ten cancer types reveals immune surveillance and immune evasion mechanisms as well as potential molecular target that could augment future immunotherapy and precision medicine strategies.