🚀 Mission View: A sharper perspective on this week's top issues that matter at the intersection of health and AI.
Much of this week’s coverage flowed out of the World Economic Forum annual meeting in Davos, where artificial intelligence was one of the many focal points. For a clear synthesis, Jeremy Kahn’s Eye on AI newsletter at Fortune captured the prevailing mood well.
One observation from Kahn’s reporting stood out. There is a growing consensus among CEOs that the bottom-up AI strategies popular two years ago, giving every employee access to tools like ChatGPT or Copilot and hoping productivity gains would organically follow, have largely run their course. Early on, leaders assumed frontline workers, closest to day-to-day processes, would naturally identify the best applications. In practice, those gains proved hard to measure and rarely translated into material improvements to either the top or bottom line.
That conclusion is at odds with much of the prevailing wisdom around responsible AI adoption, which emphasizes bringing people along, building literacy, and empowering teams to experiment safely. It also appears to conflict with recent global coordination efforts, including work from the World Economic Forum’s MINDS initiative, an AI Global Alliance led by the WEF’s Centre for AI Excellence and now encompassing more than 600 members across 500 organizations.
The MINDS paper, which was released this week and is well worth the time for anyone focused on AI adoption in their org, argues explicitly that “successful AI adoption starts with people, not technology,” and that adoption accelerates when initiatives are co-designed with employees from the start. One illustrative example is Sanofi, which involved its workforce directly in implementation. Sanofi enabled roughly 60,000 employees to co-create more than 1,300 AI use cases, embedding AI into daily work rather than treating it as a standalone technical layer. The result has been faster adoption and measurable acceleration in areas like drug discovery.
The signal from CEOs at Davos may not be that bottom-up AI adoption was wrong, but that it was incomplete. What appears to be emerging is a more hands-on leadership posture, where CEOs are taking responsibility for identifying the domains where AI can plausibly deliver material ROI, while still relying on teams closest to the work to shape how those tools are applied in practice. That also tracks with research that shows you need strong buy-in from the organizational leaders to help advance AI adoption and initiatives. So, direction comes from the top, but effectiveness still depends on frontline insight. In that sense, this is less a debate between top-down and bottom-up approaches than an acknowledgment that AI adoption now demands both: leadership willing to make hard calls about focus and investment, and organizations structured to absorb those decisions through co-creation, iteration, and trust.
🛜 Field Signals: A quick hit on this week’s industry announcements, policy developments, and ethical considerations.
🏗️ Industry
American Telemedicine Association releases AI policy framework
The American Telemedicine Association rolled out a new AI framework billed as a practical guide for providers, developers, and policymakers navigating AI's role in telehealth, virtual care, digital health, and hybrid care delivery.
Health Leadership Council releases cross-sector AI roadmap
The Health Leadership Council identifies key obstacles to broad, responsible adoption of AI across the U.S. healthcare system and provides actionable policy recommendations to address them.
OpenAI rolls out ads on ChatGPT
OpenAI introduces ads for free users; paid tiers remain ad-free. In contrast, Google by contrast says Gemini will remain ad-free, underscoring diverging incentive models.
Shadow AI spreads inside health systems
Reporting from Modern Healthcare shows unsanctioned AI tools proliferating outside IT, legal, and compliance oversight.
OpenEvidence raises at $12B valuation as physician adoption accelerates
OpenEvidence (known as “ChatGPT for Doctors”) reports use by ~40% of U.S. physicians and $100M+ in annual revenue.
Amazon launches AI health assistant for One Medical members
Not to be outdone by its AI rivals entering the health space (see announcements by OpenAI and Anthropic), Amazon embeds Health AI into One Medical, integrating AI directly into its primary care business.
FDA clears multi-condition AI triage tool for CT scans
The U.S. Food and Drug Administration has cleared Aidoc’s tool that flags up to 14 findings from one scan, signaling a regulatory evolution pertaining to AI-empowered devices, according to STAT.
🏛️ Government and Policy
FY26 Labor–HHS appropriations elevate AI workforce oversight
In the health funding bill Congress is finishing its work on, policymakers direct the Bureau of Labor Statistics to study AI labor impacts. And the bill provides $135M for NIH Office of Data Science Strategy to leverage AI, Machine Learning, and data science to accelerate the pace of biomedical innovation.
ARPA-H pushes FDA pathway for agentic AI in clinical care
The Advanced Research Projects Agency for Health is soliciting proposals to develop agentic AI assistants for use in clinical care, with the explicit goal of setting a new regulatory precedent alongside the U.S. Food and Drug Administration. The flagship effort, known as ADVOCATE (Agentic AI-Enabled Cardiovascular Care Transformation), aims to create the first FDA-authorized, patient-facing AI agent capable of providing 24/7 specialty care.
😇 Ethics and Responsible Use
Data centers, energy efficiency, and grid reality
I highly recommend this piece from DeepLearning.AI, which takes a more sober look at the energy and AI data center debate. Rather than leaning into headlines, it places data center energy use in broader context, clarifying where concerns are real, where they are overstated, and what actually matters for planning and policy.
OpenAI launches Stargate Community ‘Good Neighbor’ energy plan
Continuing the data center and energy discussion, OpenAI announced it will fund the energy infrastructure required for its data centers so local electricity prices are not pushed upward, following a similar move by Microsoft last week.
Anthropic publishes Claude’s Constitution, raising moral-status questions
Anthropic formalizes values, including instructions to disobey improper requests and acknowledges potential moral stakes. Of significant note, this is a constitution on how the model should govern itself, not necessarily how humans working at Anthropic will behave.
🔬Research and Evidence
IBM survey: AI advantage shifts from efficiency to innovation
IBM Institute for Business Value finds 64% of executives expect innovation, not cost cutting, to drive AI advantage.
Anthropic publishes fourth Economic Index based on 2M Claude chats
AI is augmenting work more than replacing it, though junior pipelines face risk.
Accenture: why AI adoption stalls before it pays off
Accenture identifies gaps in vision clarity, data quality, and role redesign.
ECRI names misuse of AI chatbots top health technology hazard for 2026
ECRI warns confident but incorrect chatbot outputs pose real patient safety risks.
🛠️ Practical Edge: Actionable tips, tools, and thoughts to help leaders strengthen capacity and apply AI in their work.
A plain-English introduction to Claude Code (even if you’re not a coder)
This 30-minute walkthrough helps demystify what all the buzz is about Anthropic’s Claude Code. It’s still fairly technical and likely too in-the-weeds for non-coders to use directly, but it’s a solid orientation for understanding how AI-native coding tools actually work and why they matter. Worth watching to build literacy, even if you never touch the keyboard.
Better meetings with Gemini
This (on-demand) session from Google shows how Gemini can be used to improve meetings end-to-end, not just generate transcripts. Use cases include real-time agenda support, summarizing decisions and open questions, tracking action items, and following up asynchronously so meetings produce momentum rather than documentation.
HBR: Why AI “workslop” happens and how to stop it
Low-quality AI output is a management failure, not a tooling failure.
HBR: How to talk to teams anxious about AI
Leaders cannot promise stability, but they can promise clarity, transparency, and repeated dialogue.
Note to my readers: I’d love to learn how you are using AI. If there’s a novel way you are deploying AI in your work, or seeing it utilized in healthcare, please feel free to shoot me a note and share: [email protected]
🌅 On the Horizon: A quick look at the developments and events expected to shape the weeks ahead.
👉 Feb. 26 - Harvard Business Review Strategy Summit: Build your transformation playbook for the AI era.
👉 Mar. 12–18, 2026 — SXSW 2026, Austin, TX
👉 Mar. 30–31, 2026 — IAPP Global Privacy Summit, Washington DC
👉 Apr. 6–9, 2026 — HumanX 2026, San Francisco, CA
👉 Apr. 10 — Ethical AI: Leadership and Governance (Virtual)
And finally, if you like what you are reading, please share this newsletter with your networks and encourage them to sign up. ✍️ 🆙 And/or, give me a shout out on LinkedIn.
Till next time,
BC
