Past AI News

All

1. Quirónsalud's in-house AI scribe: 48 Spanish EDs over 12 months, used in 45.3% of consults, 21.8% shorter on average (company-reported)

   Rare large-sample scribe deployment data, including weak results.

   @lumiaryai1 likes · 0 reposts · 0 replies · 4 viewsView post

   ExpandCollapseQuirónsalud's in-house AI scribe: 48 Spanish EDs over 12 months, used in 45.3% of consults, 21.8% shorter on average (company-reported)

   A Lumiary case write-up: Spanish hospital group Quirónsalud built an ambient AI scribe that transcribes the consult and has an LLM draft a templated report inside the EHR for the doctor to edit and validate. The Feb 2025–Jan 2026 emergency study across 48 hospitals: used in 1,032,558 consults (45.3%), monthly adoption 7.7% to 57.8%, and a mean 21.8% shorter consult. Outpatient gains were small: 15.01 vs 14.65 minutes. Primary source is an Int J Med Inform paper (doi:10.1016/j.ijmedinf.2026.106584) by group employees; retrospective and observational, so not causal.

2. NYU Langone embeds OpenEvidence's clinical AI into Epic EHR systemwide

   Clinical AI moving from individual tool to system-level EHR integration.

   @FierceHealth2 likes · 1 reposts · 0 replies · 274 viewsView post

   ExpandCollapseNYU Langone embeds OpenEvidence's clinical AI into Epic EHR systemwide

   Fierce Healthcare reports NYU Langone is embedding OpenEvidence's clinical evidence AI directly into Epic EHR workflows across its hospitals, ambulatory network and Family Health Centers. NYU Langone has 12M+ outpatient visits and 421,000 ED visits a year and 6,000+ physicians and faculty. OpenEvidence says about two-thirds of U.S. physicians use it and verified clinicians ran 45M queries in October (company figures). It follows NewYork-Presbyterian and Mount Sinai deployments in New York.

3. OpenAI Devs: immunologist Derya Unutmaz uses Codex to simulate single-cell drug response

   A case of coding agents entering life-science research.

   @OpenAIDevs602 likes · 45 reposts · 79 replies · 57,618 viewsView post

   ExpandCollapseOpenAI Devs: immunologist Derya Unutmaz uses Codex to simulate single-cell drug response

   OpenAI's developer account features immunologist Derya Unutmaz using Codex to simulate how a single cell might respond to a drug. He is one of nine community builders in OpenAI's Codex Originals series (openai.com/codex-originals). No accuracy or validation data is given; it is a showcase.

4. Nature Reviews Drug Discovery Perspective: where AI in drug discovery stands and how to raise translational relevance

   A review pharma can use when weighing AI R&D spend.

   @NatRevDrugDisc42 likes · 14 reposts · 0 replies · 3,052 viewsView post

   ExpandCollapseNature Reviews Drug Discovery Perspective: where AI in drug discovery stands and how to raise translational relevance

   Nature Reviews Drug Discovery's November issue carries a Perspective, 'Artificial intelligence in drug discovery — what it is, where we stand and the path forward', reviewing the field and recommending ways to increase translational relevance. Full text at nature.com/articles/s41573-026-01496-2, with a free share link. The post does not list the specific recommendations.