Next‑Gen AI, policy making and the future of radiology practice are on the table — and the ACR’s Data Science Institute® (DSI) is on your side
This is a momentous time for our profession — and I don’t use that term lightly. AI technology is no longer a "future" consideration, but increasingly a present-day reality in imaging care. For many, this transformation cannot come too soon, provided it works. We are all feeling the strain of an ever-widening gap between the skyrocketing demand for imaging services and our collective ability to provide high-quality, timely care. Traditional solutions are falling short, because in times of severe workforce shortage, we simply cannot adequately scale our current human-based way of practicing.
Looking back at the past six years of working with phenomenal volunteer leaders and ACR staff within the IT Commission and the DSI, I am filled with gratitude. We have moved from "PowerPoint to Practice," turning high-level AI concepts into tangible programs and infrastructure to support member practices. Our team has created AI Central, conducted the joint ACR-RSNA SERT effort (Safe, Effective, Reliable and Transparent AI), developed the ARCH-AI recognition program and the ASSESS-AI registry, which is rapidly gaining traction. These aren’t just acronyms; they are actual building blocks of a community of practices who are taking charge of implementing AI responsibly. At the upcoming Annual Meeting, the ACR Council will also consider the ratification of the first-ever Practice Parameter in Imaging AI, alongside consideration for the development of a concept for a matching accreditation program.
While the first generation of Software as a Medical Device (SaMD) often fell short of the marketing hype — offering limited value propositions like simple triage — we are now entering a new chapter with more powerful next-gen technology on the horizon.
This year is shaping up to be the year of the Foundation Model in Radiology. In June 2025, a pioneering team at Northwestern University published groundbreaking insights into applying a chest X-ray foundation model for report drafting. This was a clarion call: Radiology specific generative AI was entering real-world clinical practice. Increasingly, we are learning about practices implementing report-drafting radiology foundation models under IRB oversight. Of course, this means that many practicing radiologists will need to develop the skill to efficiently proofread AI drafted reports — which can be compared to working with a senior resident (in the best-case scenario). We will need to examine if these next-generation tools finally offer the promise of comprehensive support for our professional work and greater effectiveness — a promise that the early iterations of AI never quite delivered.
We will need to examine if these next-generation tools finally offer the promise of comprehensive support for our professional work and greater effectiveness — a promise that the early iterations of AI never quite delivered.
We are currently witnessing another fascinating, albeit slightly “chaotic,” phenomenon. Generative AI innovation is quickly outpacing traditional regulatory and clinical guideline lifecycles. Practices are pulling ahead of existing regulatory frameworks, though under strict local oversight. However, the regulatory landscape may be slowly shifting to meet us.
There are indicators that the federal government is looking to deregulate and remove perceived “burdens” to Ai innovation. A pending citizen’s petition filed by an AI company seeks a "regulatory easement" for SaMD, requesting exemption from pre-market 510(k) clearance under specific circumstances. What is the trade-off if this request is accepted? Companies would have to commit to preconceived post-deployment monitoring approaches that are subject to discretionary FDA audit. This means the actual behavior of models must be tracked in our real-world practices and much of this quality assurance activity falls on the radiologists who deploy them.
This is where you come in. Vendors will need to partner with ACR member practices to achieve this expectation, and the ACR ASSESS-AI registry, a core component of our ARCH-AI quality assurance program, offers a viable solution for this type of monitoring by an independent trusted third party, the ACR. A growing number of radiology use cases have been developed for the registry and others await “site-triggered” go live. Research and development are underway in the DSI to expand this capability to help get our arms around understanding the behavior of multi-channel AI and radiology report drafting foundation models in the radiology specific context.
To help members better understand the current and evolving regulatory landscape, I am thrilled to announce that we will welcome the FDA leadership team from the Office of Radiological Health to a "Lunch and Learn" session at the 2026 ACR Annual Meeting. This is a rare opportunity to engage in an interactive, live conversation to better understand how SaMD regulation is currently designed and evolving. Mark your calendars!
Speed of adoption and risk tolerance varies across the globe — countries with an even wider gap between supply and demand for radiologists’ services, may be taking risks in AI implementation sooner and on a larger scale than the US. ACR is committed to fostering an international dialogue to facilitate bilateral learning from successes and opportunities. We are now offering a low-cost, limited scope international ACR Informatics membership and are expanding the ARCH-AI recognition to sites outside the U.S. for the first time. In learning from each other, we can better ensure that the human-machine partnership remains focused on optimal patient care informed by real world experiences.
What can you do right now?
Learning to navigate this new era infused with AI is very much like learning to ride a bicycle: it is nearly impossible to understand the balance until you actually sit on the seat and start pedaling. I invite you to get on the bike with us.
Bolstering Awareness in Wisconsin
The WRS has taken steps to improve patient care by uniting breast radiologists to ignite positive change within the specialty.
Read more
A New Era for the Commission on Economics
Outgoing Chair Gregory N. Nicola, MD, FACR, reflects on recent wins — and his hopes for the future of the Commission and the specialty.
Read more
Workforce Economics
The state of the radiology workforce has affected many practices across the country, but there are ways to navigate these challenging times.
Read more
