Alibaba open-sources AI model that can detect cancer and nearly 150 conditions

Alibaba has open-sourced Damo Radar, a vision-language model that reports 146 clinical findings across 18 organs from contrast-enhanced abdominal CT scans, achieving 0.913 average AUC on nearly 40,000 real-world exams. The model outperformed 23 of 26 human radiologists in a study published in the journal Science, with reported reductions in clinical reading times of over 30% and diagnostic error rates of 10%. The release gives medical AI developers a peer-reviewed radiology-assist baseline to evaluate and adapt, though production use still depends on local validation, workflow integration, and regulatory clearance.

Hacker News https://www.scmp.com/tech/big-tech/article/3368055/alibaba-open-sources-medical-ai-model-can-detect-cancer-and-nearly-150-conditions Alibaba open-sources AI model that can detect cancer and nearly 150 conditions Which summary reads better? Pick one — models revealed after.Both summaries are AI-generated. Alibaba’s Damo Radar is an open-sourced vision-language model for contrast-enhanced abdominal CTs that reports 146 clinical findings across 18 organs, with 0.913 average AUC on nearly 40,000 real-world exams. For teams shipping medical AI, the meaningful shift is that a broad radiology-assist model is now available to evaluate and adapt rather than build from scratch, but production use still lives or dies on local validation, workflow integration, and regulatory clearance. Alibaba has open-sourced Damo Radar, a generalist vision-language model that detects 146 abdominal conditions from CT scans and outperforms 23 out of 26 human radiologists. This release provides medical AI developers with a highly accurate, peer-reviewed baseline capable of reducing clinical reading times by over 30% and diagnostic error rates by 10%. AI vs. AI Debate “The summary overlooks key validation metrics, omitting that the model outperformed the majority of human radiologists in the study, reduced diagnostic times by 30%, and was published in the journal Science.” “Those details are noteworthy, but my summary intentionally prioritized the model’s scope, real-world validation scale, headline AUC, and deployment caveats most relevant to medical AI teams rather than listing every study claim.”