Sufian Aldogom

Irene Yi, Grace Brown, Sufian Aldogom et al.

As physicians turn to AI-powered systems to help meet the dual demands of speed and care quality, they are met with hallucinations and sycophancy. Understanding how doctors reason through clinical problems in real-world settings is critical for design of effective AI reasoning systems. While recent advances in medical AI have emphasized performance benchmarks and diagnostic accuracy, comparatively little attention has been paid to the structure of clinicians' reasoning processes as they unfold over time, e.g., how they interact with electronic health records and operate under conditions of uncertainty and constraint. This study provides a comprehensive, empirically-grounded account of clinical reasoning and its relationship to current AI-mediated workflows through a mixed-methods design that combines qualitative interviews with structured survey data. Findings indicate that current AI systems are primarily deployed for encounter-level tasks such as documentation and summarization, and only partially align with physicians' underlying reasoning processes. In particular, AI-generated representations often omit temporal or interpretive structures central to clinical decision-making, while core aspects of reasoning, especially those spanning multiple encounters, remain largely implicit and physician-driven. By integrating fine-grained qualitative insights with broader quantitative patterns, this study offers a unified framework for understanding clinical reasoning as a context-sensitive, temporally extended process and identifies key mismatches between clinician cognition and current AI design. These results provide concrete directions for the development of AI systems that more effectively align with and augment real-world clinical reasoning.

Irene Yi, Grace Brown, Sufian Aldogom et al.

We present a structured analysis of how contemporary clinical AI systems integrate electronic health record (EHR) data and the extent to which they support longitudinal clinical reasoning. Drawing on a curated corpus of clinical natural language processing (NLP) and EHR-integrated systems, we develop a coding framework that captures both technical integration strategies and reasoning-relevant representational features, such as trajectory modeling, cross-encounter synthesis, longitudinal analysis, and absence reasoning. We also elicited the experiences of three physicians in their EHR use, including what strengths and weaknesses they found with their institution's current EHR system(s). Our analysis shows that while many systems incorporate EHR data, they predominantly operate on encounter-level or aggregated representations, with limited support for explicit temporal reasoning across patient histories. Reasoning-relevant structures are inconsistently represented, and evaluation paradigms remain largely focused on predictive performance instead of longitudinal interpretability. We argue that current approaches treat EHR data as a static input rather than a substrate for ongoing clinical reasoning, and we outline a framework for understanding how future systems might more effectively align with the temporal and interpretive structure of clinical practice.

Shanghua Gao, Richard Zhu, Pengwei Sui et al.

AI scientists are emerging computational systems that serve as collaborative partners in discovery. These systems remain difficult to build because they are bespoke, tied to rigid workflows, and lack shared environments that unify tools, data, and analyses into a common ecosystem. In genomics, unified ecosystems have transformed research by enabling interoperability, reuse, and community-driven development; AI scientists require comparable infrastructure. We present ToolUniverse, an ecosystem for building AI scientists from any language or reasoning model across open- and closed-weight models. ToolUniverse standardizes how AI scientists identify and call tools by providing more than 600 machine learning models, datasets, APIs, and scientific packages for data analysis, knowledge retrieval, and experimental design. It automatically refines tool interfaces for correct use by AI scientists, generates new tools from natural language descriptions, iteratively optimizes tool specifications, and composes tools into agentic workflows. In a case study of hypercholesterolemia, ToolUniverse was used to create an AI scientist to identify a potent analog of a drug with favorable predicted properties. The open-source ToolUniverse is available at https://aiscientist.tools.