The Royal College of Radiologists (RCR), the Institute of Physics and Engineering in Medicine (IPEM), and the Society of Radiographers (SoR) have called for a properly trained and funded workforce and clear, consistent regulations for AI implementation.
This plea came as part of their responses to the Medicines and Healthcare products Regulatory Agency (MHRA) consultation document on the regulation of AI in healthcare.
Together, they developed the following three recommendations toward regulatory priorities:
- End-to-end assurance across the AI lifecycle. This includes regulations requiring proportionate premarket evidence, transparent communication of limitations, and mandatory postmarket surveillance to detect performance drift and bias. Clinicians would retain oversight throughout this process.
- Workforce capacity as a patient safety requirement. National workforce planning, funded training pathways, recognized roles, and protected time must be central to regulation via a trained, resourced workforce.
- Clear system-wide accountability. Regulation should be clear on where responsibility lies between manufacturers, healthcare organizations, and professionals. This includes expectations for transparency, training, postmarket monitoring, and liability.
The societies stated that AI's use is expanding rapidly as the technology is embedded across imaging and radiotherapy. They called for regulation to be grounded in real clinical practice, reflecting patient safety, workforce capacity, and National Health Service (NHS) delivery realities.
Also, the RCR has responded to the U.K.'s AI and robotic pilot scheme to detect lung cancer. Go to the RCR website for more details.


![A normal mammogram confirmed by three-year radiologic follow-up illustrates reader-marked regions of interest (ROIs) during (A) unaided (round 1) and (B) artificial intelligence (AI)–assisted (round 2) reading. Each colored dot represents an ROI for recall by a human reader. Readers could mark more than one ROI per case, represented by multiple dots of the same color. During AI-assisted reading, the AI system displayed three visible prompts: two with suspicion of malignancy scores of 35% (left mediolateral oblique [L MLO] and craniocaudal [L CC]) and one with a suspicion of malignancy score of 10% (right craniocaudal [R CC]), shown as polygonal overlays. Without AI, six of 10 readers (60%) marked a false-positive ROI. With AI assistance, this fell to two of 10 (20%). R MLO = right mediolateral oblique.](https://img.auntminnieeurope.com/mindful/smg/workspaces/default/uploads/2026/07/2026-07-14-radiology-mammogram-ai-auto-bias.H0bYO8QlWs.jpg?auto=format%2Ccompress&fit=crop&h=100&q=70&w=100)







![A normal mammogram confirmed by three-year radiologic follow-up illustrates reader-marked regions of interest (ROIs) during (A) unaided (round 1) and (B) artificial intelligence (AI)–assisted (round 2) reading. Each colored dot represents an ROI for recall by a human reader. Readers could mark more than one ROI per case, represented by multiple dots of the same color. During AI-assisted reading, the AI system displayed three visible prompts: two with suspicion of malignancy scores of 35% (left mediolateral oblique [L MLO] and craniocaudal [L CC]) and one with a suspicion of malignancy score of 10% (right craniocaudal [R CC]), shown as polygonal overlays. Without AI, six of 10 readers (60%) marked a false-positive ROI. With AI assistance, this fell to two of 10 (20%). R MLO = right mediolateral oblique.](https://img.auntminnieeurope.com/mindful/smg/workspaces/default/uploads/2026/07/2026-07-14-radiology-mammogram-ai-auto-bias.H0bYO8QlWs.jpg?auto=format%2Ccompress&fit=crop&h=112&q=70&w=112)








