---
title: "OpenEvidence CEO tells medical students to take paper exams as OpenAI and Anthropic give doctors AI free"
description: "Daniel Nadler built a $15 billion company on giving doctors AI answers free. On Monday he warned the doctors still training on his tool that AI erodes the skill that makes an answer worth having."
publisher: "The Inference"
section: "Health"
published: 2026-11-18T17:17:54.243Z
modified: 2026-11-18T17:17:54.243Z
canonical: https://theinference.org/article/openevidence-ceo-tells-medical-students-to-take-paper-exams-as-openai-and-anthropic-give-doctors-ai-free
language: en
keywords: "OpenEvidence, Medical AI, OpenAI, Anthropic, Apple"
---

# OpenEvidence CEO tells medical students to take paper exams as OpenAI and Anthropic give doctors AI free

> Daniel Nadler built a $15 billion company on giving doctors AI answers free. On Monday he warned the doctors still training on his tool that AI erodes the skill that makes an answer worth having.

## Paper exams, from the CEO who put AI in the chart

Daniel Nadler, founder and chief executive of the medical-literature search engine [OpenEvidence](https://theinference.org/markets/companies/openevidence), [told Semafor's Silicon Valley & The World event on November 16](https://thenextweb.com/news/medical-students-should-take-paper-exams-openevidence-ceo-tells-semafor) that medical students should "learn the hard way, do paper exams" even if they also use AI [1]. "The risk of cognitive offloading is really, really acute," he said, cognitive offloading being the handing of mental work to a machine, and sharpest for students who have studied with AI from the start [1]. Semafor put his company's value at $15 billion [1], a mark set by [a $250 million September round](https://theinference.org/article/openevidence-raised-250-million-at-15-billion-pro-medicus-fell-48-in-a-year-on-the-same-doctor-screen), earned on advertising revenue, with verified doctors paying nothing [2].

*Chart: **OpenEvidence's valuation climbed from $1 billion to $15 billion in 19 months** Column chart of OpenEvidence's post-money valuation across five funding rounds from February 2025 to September 2026, rising from $1 billion to $15 billion.*

*Post-money valuation by funding round, February 2025 to September 2026; compiled from round reporting collected on Wikipedia (CNBC, NYT, Forbes, Fierce Healthcare). [2]*

## Two small studies back the warning

The warning has numbers behind it. A 2025 Lancet study found skilled doctors spotted fewer precancerous growths without AI after months of using it, 22.4% unaided against 28.4% before [1]. An MIT Media Lab study of 54 people found the weakest brain connectivity among those writing essays with ChatGPT [1]. Both were small [1].

The bite is the deployment: [since October 20, OpenEvidence has sat inside NYU Langone Health's Epic medical-record system](https://i10x.ai/news/openevidence-joins-nyu-langone-epic-workflow), reaching 12.4 million outpatient visits a year and more than 6,000 physicians, among them the 2,000-plus residents and fellows training at NYU's two medical schools [3]. The cohort Nadler wants on paper already has his tool in the chart.

## Free answers from three giants

The same week stacked free supply against that valuation. [Anthropic](https://theinference.org/markets/companies/anthropic), maker of Claude, [agreed in September to supply OpenEvidence free to clinicians in about 100 countries](https://www.ibtimes.com/anthropic-taking-medical-ai-global-new-partnership-it-aims-take-work-some-100-countries-3807778), with financial terms undisclosed [4].

On November 15, [Apple](https://theinference.org/markets/companies/apple) [put OpenEvidence in front of up to 200,000 U.S. consumers in a health study](https://www.healthcaredive.com/news/apple-openevidence-conversational-health-ai-study-app/874714/), no commercial terms disclosed [5], a launch [The Inference covered on November 16](https://theinference.org/article/apple-s-200-000-person-health-ai-study-runs-on-openevidence-with-no-disclosed-price). A day later, [OpenAI](https://theinference.org/markets/companies/openai) [loaded more than 20 health plugins into ChatGPT for Healthcare](https://hitconsultant.net/2026/11/16/openai-launches-healthcare-computer-use-agents-clinical-search-20-plugins/), where ChatGPT for Clinicians is free for eligible U.S. doctors, and priced its GPT-6 Luna admin agent at 8.5 cents a task, 98% below the frontier model [6][7].

## Licensable content is what free answers cannot copy

The defense is scale and licenses. U.S. clinicians consulted OpenEvidence 42 million times in August alone [4], and the company [holds content agreements with NEJM, JAMA, Nature, NCCN and Cochrane](https://www.openevidence.com/about) [8]. Licensed clinical content can hold a price where general answers cannot; advertising aimed at doctor screens is what free supply cuts. Before paying the September mark, investors should want the Apple and health-system economics disclosed, and a view on what a pharma ad is worth beside a free ChatGPT. Nadler's own warning prices the risk: the scarce asset is not the answer but the doctor who can check it.

## Takeaway

Daniel Nadler says the risk of handing mental work to a machine is acute for the doctors training on his own tool, which already sits inside NYU Langone's medical records. The same week, Anthropic, Apple and OpenAI all put medical AI in front of doctors for free, leaving a $15 billion valuation resting on pharma advertising and licensed literature.

## Sources

1. [The Next Web, "Medical students should take paper exams, OpenEvidence CEO tells Semafor", 17 November 2026](https://thenextweb.com/news/medical-students-should-take-paper-exams-openevidence-ceo-tells-semafor)
2. [Wikipedia, "OpenEvidence", retrieved 18 November 2026](https://en.wikipedia.org/wiki/OpenEvidence)
3. [i10X News, "OpenEvidence Joins NYU Langone Epic Workflow", October 2026](https://i10x.ai/news/openevidence-joins-nyu-langone-epic-workflow)
4. [International Business Times, "Anthropic Is Taking Medical AI Global With a New Partnership", 23 September 2026](https://www.ibtimes.com/anthropic-taking-medical-ai-global-new-partnership-it-aims-take-work-some-100-countries-3807778)
5. [Healthcare Dive, "Apple, OpenEvidence launch conversational health AI study and app", 16 November 2026](https://www.healthcaredive.com/news/apple-openevidence-conversational-health-ai-study-app/874714/)
6. [HIT Consultant, "HLTH26: OpenAI Expands ChatGPT for Healthcare: Computer-Use Agents, 99% Clinical Safety Benchmarks, and 20+ Integrated Health Tech Plugins", 16 November 2026](https://hitconsultant.net/2026/11/16/openai-launches-healthcare-computer-use-agents-clinical-search-20-plugins/)
7. [Tecnología Geek, "ChatGPT for Healthcare estrena agentes de computadora", November 2026](https://tecnologiageek.com/chatgpt-for-healthcare-agentes-de-computadora/)
8. [OpenEvidence, "About OpenEvidence", retrieved 18 November 2026](https://www.openevidence.com/about)


---

OpenEvidence CEO tells medical students to take paper exams as OpenAI and Anthropic give doctors AI free — The Inference. Canonical: https://theinference.org/article/openevidence-ceo-tells-medical-students-to-take-paper-exams-as-openai-and-anthropic-give-doctors-ai-free
