What is one specific practice or element that is essential for scaling AI solutions in healthcare while maintaining trust and safety?

Most answers emphasize human oversight, governance, and continuous validation as essential for scaling AI in healthcare, with some highlighting transparency, workflow integration, and technical safeguards.

22 answers

Answers

  • Ashley Alexander

    VP, Health Products · OpenAI

    Clinician Red Teaming

    Ashley Alexander says her organization has hundreds of clinicians red teaming and evaluating every major step of training that involves health context, which she describes as part of meeting responsibility and safety standards.

    Peter Yang · Watch on YouTube
  • Andreas Cleve

    Co-founder & CEO · Corti

    Governable, Compliant AI

    Andreas Cleve says healthcare's depth and unmet needs require "really governable, compliant AI," which he says general-purpose models never achieved, so someone had to focus on it full-time.

    The European VC · Watch on YouTube
  • Chris Carmody

    Chief Information Officer · Upmc

    Governance

    Chris Carmody says scaling AI responsibly starts with governance and the processes around it, which he says are needed to identify and focus on the top problems so efforts can succeed and build from there.

    The answer starts at 7:04.

    Becker’s Healthcare Podcast · Listen on Apple Podcasts
  • Phoebe L. Yang

    Lecturer in Law · Stanford Law School

    Physician Review

    Phoebe L. Yang says physician review of AI outputs is essential, citing Doximity's practice of enlisting over 12,000 physicians to review every clinical answer before it reaches clinicians.

    Becker Private Equity & Business · Becker Private Equity & Business · Listen at 12:01
  • Ilya Burkov

    Global Head of Scientific AI & Healthcare · Nebius

    Human-in-the-Loop

    Ilya Burkov says healthcare is the best case for AI because it already has guardrails like licensing boards and malpractice law, and he describes humans being in the loop all the time, with AI tools acting as assistants rather than replacements.

    Second Opinion with Christina Farr · Watch on YouTube
  • Hal Wolf

    President and Chief Executive Officer · Himss

    Governance

    Hal Wolf says governance is one of the things that must be in place, along with training, processes and impact measurement, for organizations to take advantage of technology.

    The answer starts at 6:50.

    The Big Unlock · Listen on Apple Podcasts
  • Nirav Shah, MD, MPH

    Associate Chief Medical Informatics Officer, AI and Innovation · Endeavor Health

    Continuous Model Validation

    Nirav Shah says a partner must remain accountable for continuous model validation, because AI models drift as clinical populations and workflows evolve after go-live.

    The answer starts at 7:07.

    Becker’s Healthcare Podcast · Listen on Apple Podcasts
  • Katherine Adams, MS

    Director, Clinic Operations, Primary Care · Commonspirit Health

    Ongoing Partnership

    Katherine Adams says she needs her technology partners' ongoing partnership after go-live so AI tools keep learning, outcomes improve, and hallucinations are avoided, calling for the same commitment to patient safety.

    The answer starts at 5:20.

    Becker’s Healthcare Podcast · Listen on Apple Podcasts
  • Zeke Emanuel

    Vice Provost for Global Initiatives · University Of Pennsylvania

    Physician Oversight

    Zeke Emanuel says the real disagreement over autonomous AI is whether it can manage a patient without case-by-case physician oversight, which he frames as the defining issue of trust and safety.

    AI and Healthcare · AI and Healthcare · Listen at 13:13
  • Colin LeClair

    Chief Executive Officer · Connections Health Solutions

    Building Safeguards into Contracts

    Colin LeClair says his team builds safeguards into their contracts, so that if the vendor company folds they can absorb the technology and keep operating the solutions they've built.

    The answer starts at 13:46.

    Becker’s Healthcare Podcast · Listen on Apple Podcasts
  • Amit Malik

    Founder & CEO · Amaha

    Human Accountability

    Amit Malik says what is and always will be missing from an LLM or AI chatbot is the human and clinical accountability angle, which he identifies as essential.

    The answer starts at 19:14.

    Say Again?! With Amrita Tripathi · Listen on Apple Podcasts
  • David Niewolny

    Sr. Dir, Global Head of Business Development - Healthcare / Medical · NVIDIA

    Clear Human Decision Point

    David Niewolny says governance should include a clear human decision point in AI applications, designed to give clinicians superpowers rather than remove decisions from them.

    The answer starts at 35:44.

    Faces of Digital Health · Listen on Apple Podcasts
  • Don Woodlock

    President · Intersystems

    Human-in-the-Loop

    Don Woodlock says human-in-the-loop is the most obvious essential, arguing AI must stay human led because patients and clinicians want human expertise and make the final choices.

    The answer starts at 5:33.

    The Big Unlock · Listen on Apple Podcasts
  • Munjal Shah

    Senior Product Manager, Wearables · Google

    Using Multiple Supervised Models

    Munjal Shah says the way to make AI safe is to use a whole host of models working together, noting his team uses 31 models—one main model supervised by 30 others—rather than a single model.

    The answer starts at 7:07.

    "Second Opinion" with Christina Farr, Ash Zenooz MD & Luba Greenwood JD · Listen on Apple Podcasts
  • Peter Lee

    President, Microsoft Science · Microsoft

    Integrate AI into Workflow

    Peter Lee says the answer is workflow: integrating AI into clinical workflow in a highly reliable and consistent way is how outcomes are driven, rather than finding the perfect model.

    Second Opinion with Christina Farr · Watch on YouTube
  • Rhett Alden

    CTO, Elsevier Health Markets · Elsevier

    Robust Validation Framework

    Rhett Alden says his team has a robust validation framework for CKAI and publishes results on errors and hallucination risk, which he says provides grounding for clinicians.

    The answer starts at 3:59.

    The AI in Business Podcast · Listen on Apple Podcasts
  • Anirban Maitra

    Director, Perlmutter Cancer Center · Nyu Langone Health

    Qualified Physician Oversight

    Anirban Maitra says the most important thing is using AI safely, which means having qualified physicians analyze and check the results rather than letting AI diagnose without oversight.

    The answer starts at 10:15.

    Becker’s Healthcare Podcast · Listen on Apple Podcasts
  • Hannah Galvin, MD, FAAP, FAMIA, CHCIO

    Chief Health Information Officer · Cambridge Health Alliance

    AI Governance

    Hannah Galvin says her organization was an early adopter of mature AI governance, which they have evolved over time to keep up with industry trends and the growing number of tools and opportunities available to them.

    The answer starts at 7:37.

    Becker’s Healthcare Podcast · Listen on Apple Podcasts
  • Albert Villarin, MD MBA FACEP

    VP | Chief Medical Informatics Officer · Northwell Health

    Training Clinicians

    Albert Villarin says training clinicians is essential, noting his organization trains residents through a clinical informatics rotation that brings AI into their education, and that it is healthcare's responsibility to train them to use tools like AI.

    Becker’s Healthcare Podcast · Becker’s Healthcare Podcast · Listen at 18:24
  • Kimberly Powell

    Vice President / General Manager, Healthcare at NVIDIA · Nvidia

    Keep Clinicians in the Loop

    Kimberly Powell says it is so important to have health care professionals in the loop, both when using the tools and through clinicians building their domain knowledge into the systems, so the AI can serve as a second opinion.

    Second Opinion with Christina Farr · Watch on YouTube
  • Julian Tang

    COO for Innovation Office · BlackRock

    Transparency

    Julian Tang says transparency is essential for scaling AI, because when AI becomes transparent and people can see why it produced an answer, they start to trust it more.

    The AI in Business Podcast · The AI in Business Podcast · Listen at 11:59
  • Muralidhar Vemulapalli

    Chief Enterprise Architect/ Acting CTO · Arkansas Blue Cross And Blue Shield

    Semantic Layer

    Muralidhar Vemulapalli says the enterprise semantic layer is the most important element, because machines, unlike developers, don't know what column names mean, so defining that meaning is essential.

    CDO Magazine Podcast Series · CDO Magazine Podcast Series · Listen at 4:11

More HealthTech questions

View all →