Harrison.ai is highlighting preliminary results of an independent healthcare AI challenge conducted by Mass General Brigham and the American College of Radiology (ACR) that included the company's radiology-specific foundation model, Harrison.rad.1.
Mass General Brigham and the ACR's Data Science Institute organized the challenge, which included 113 radiologists performing 2,840 blinded evaluations across 117 reports during the ACR's annual meeting.
Harrison.rad.1 was released last year. In the AI challenge, it outperformed models from OpenAI, Anthropic, and Google on VQA-Rad, a widely used benchmark for evaluating and comparing the performance of multimodal foundational models on medical tasks, with an accuracy and precision rate of 82% on closed questions filtered for plain x-rays, the company said.
![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.auntminnie.com/mindful/smg/workspaces/default/uploads/2026/07/2026-07-14-radiology-mammogram-ai-auto-bias.H0bYO8QlWs.jpg?auto=format%2Ccompress&dpr=2&fit=crop&h=167&q=70&w=250)
