![\"Artificial](\"https:\/\/rtmedical.com.br\/wp-content\/uploads\/2026\/03\/acr-legislacao-ia-radiologia.jpg\")
While not a new trend, the legislative push gained significant momentum in 2025 and 2026. Multiple states have introduced bills addressing transparency requirements \u2014 such as informing patients when AI algorithms participate in analyzing their imaging studies \u2014 mandatory physician oversight, clinical validation standards, and liability frameworks for AI-related diagnostic errors.<\/p>\n
What State Legislation Proposes<\/h2>\n
Although details vary across states, several recurring themes emerge in the bills the ACR monitors. Algorithmic transparency<\/strong> is a common pillar: multiple states want patients notified when AI tools participate in the analysis of their imaging exams. The reasoning is that patients have a right to know whether their mammogram, CT scan, or X-ray was interpreted with computational assistance.<\/p>\n Another central theme is mandatory physician oversight<\/strong>. Most proposals require a licensed radiologist to review and approve any AI-generated result before it is communicated to the patient or entered into the medical record. This requirement aligns with the ACR’s longstanding position that AI should serve as a support tool for radiologists, never as a replacement.<\/p>\n Liability<\/strong> questions also appear frequently. Who is responsible when an AI algorithm produces an incorrect diagnosis? The software developer? The hospital that acquired it? The radiologist who relied on the result? These questions lack uniform answers, and each state is seeking to define its own regulatory framework.<\/p>\n The ACR has advocated for a balanced approach that promotes innovation without compromising patient safety. The organization argues that overly restrictive regulations could inhibit the development and adoption of tools that demonstrably improve diagnostic efficiency. At the same time, it recognizes the need for clear safeguards.<\/p>\n Recent data on AI’s impact in radiology practice are compelling. Studies show that AI tools can reduce mammography workload by up to 64%<\/a>, while computer-aided detection algorithms are already in widespread use for lung cancer screening and brain imaging analysis. The regulatory question, therefore, is not whether AI will be used in radiology \u2014 it already is \u2014 but under what conditions and with what guarantees.<\/p>\n In the United States, regulation of AI-based medical devices occurs primarily at the federal level through the FDA, which has already cleared more than 900 AI\/ML devices for clinical use \u2014 the majority in radiology. However, the FDA regulates the device itself, not necessarily how it is implemented in clinical practice. State legislation aims to fill that gap.<\/p>\n Regulatory fragmentation across states could create challenges for manufacturers and hospitals operating in multiple jurisdictions. An FDA-cleared AI algorithm might be freely deployable in one state but require mandatory patient notification in another and real-time radiologist oversight in a third. The ACR has been vocal about the need for some level of harmonization to prevent an unworkable regulatory patchwork.<\/p>\n The U.S. legislative wave is not occurring in isolation. The European Union is advancing the AI Act, which classifies healthcare AI applications as “high-risk” and imposes stringent compliance requirements. Several other nations are developing their own regulatory frameworks for AI-based medical software. The ECR 2026 highlighted AI and screening as central trends<\/a>, reinforcing that the regulatory landscape will be inseparable from radiology practice in coming years.<\/p>\n For practicing radiologists worldwide, monitoring the U.S. state-level regulatory landscape is strategic. Decisions made in the United States and Europe frequently shape the regulations that eventually emerge in other markets, including Latin America and Asia-Pacific.<\/p>\n The most optimistic view is that well-designed regulation could actually accelerate AI adoption in radiology by increasing physician and patient confidence. Quality certifications, transparency requirements, and clear liability frameworks can differentiate serious tools from questionable ones. The challenge lies in finding the balance between protection and innovation \u2014 a challenge that U.S. state legislators, with the ACR’s close engagement, are only beginning to address.<\/p>\nACR’s Position and Practice Impact<\/h2>\n
Federal vs. State Regulatory Landscape<\/h2>\n
Global Context and International Trends<\/h2>\n
Outlook: Regulation as an Innovation Enabler<\/h2>\n