− LIFESTYLE Precision nutrition on a global scale necessitates an understanding of food not as static collections of so-c…+ SCIENCE Impedance flow cytometry (IFC) is a label-free, single-cell measurement technique that captures biophysical p…+ HEALTH Machine learning is increasingly reshaping psychological and mental health research by complementing traditio…− HEALTH Generative artificial intelligence in hematology is entering a new phase. The dominant question, whether larg…− CRIME Background Percutaneous coronary intervention effectively treats flow-limiting stenoses; however, untreated n…− HEALTH Objective To evaluate the real-world diagnostic performance of a multimodal large language model (LLM) for im…− HEALTH Artificial intelligence (AI) is increasingly present in dental education and clinical practice through system…+ HEALTH The diagnosis of major depressive disorder (MDD) currently relies on subjective clinical assessment, undersco…
TruaceTracing the truth around AITuesday, September 22, 2026
Subscribe
Health·P Space·Evidence-backed problem·Published 2026-09-22

From Answers to Agents: What Must Change Before Generative and Agentic AI Become Clinical Infrastructure in Hematology

Abstract: Generative artificial intelligence in hematology is entering a new phase. The dominant question, whether large language models are accurate enough for clinical decision support, is being overtaken by a harder one, as systems shift from answering questions to acting: extracting structured cases, routing them, classifying variants, and grounding recommendations in guidelines and case memory. In 2026, a hematology agent achieved roughly 83% concordance with tumor-board decisions in a prospective silent trial, with…

TRV-2026-1168Peer-reviewedPermanent record — cite & verify
From Answers to Agents: What Must Change Before Generative and Agentic AI Become Clinical Infrastructure in Hematology

Do you think you will have to move because of climate change? by European Investment Bank. CC BY-SA 3.0 igo · https://creativecommons.org/licenses/by-sa/3.0/igo/deed.en

The quick read

Generative artificial intelligence in hematology is entering a new phase. The dominant question, whether large language models are accurate enough for clinical decision support, is being overtaken by a harder one, as systems shift from answering questions to acting: extracting structured cases, routing them, classifying variants, and grounding recommendations in guidelines and case memory.

In 2026, a hematology agent achieved roughly 83% concordance with tumor-board decisions in a prospective silent trial, with hallucinations in 0.3%, suggesting that for well-structured tasks the binding constraint is shifting from accuracy to governability. For hematologists, this reframes AI adoption as a governance and human-factors program rather than a search for a better model.

Main points
  • Generative artificial intelligence in hematology is entering a new phase.
  • The dominant question, whether large language models are accurate enough for clinical decision support, is being overtaken by a harder one, as systems shift from answering questions to acting: extracting structured cases, routing them, classifying variants, and grounding recommendations in guidelines and case memory.
  • In 2026, a hematology agent achieved roughly 83% concordance with tumor-board decisions in a prospective silent trial, with hallucinations in 0.3%, suggesting that for well-structured tasks the binding constraint is shifting from accuracy to governability.
Problem

From Answers to Agents: What Must Change Before Generative and Agentic AI Become Clinical Infrastructure in Hematology: In 2026, a hematology agent achieved roughly 83% concordance with tumor-board decisions in a prospective silent trial, with hallucinations in 0.3%, suggesting that for well-structured tasks the binding constraint is shifting from accuracy to governability.

The rundown

In 2026, a hematology agent achieved roughly 83% concordance with tumor-board decisions in a prospective silent trial, with hallucinations in 0.3%, suggesting that for well-structured tasks the binding constraint is shifting from accuracy to governability. Moving from experimental support to reliable infrastructure requires four changes: validating trajectories rather than answers; allocating autonomy inversely to case complexity; treating deployed models as regulated instruments under continuous surveillance; and engineering against automation bias, with data-privacy and consent safeguards throughout.

Sources

Reader signal

How should this claim be treated?

The debate