Video summary
This video shows how Northwestern Medicine is using AI to move healthcare from reactive treatment toward more proactive care. The organization is working with Dell Innovation Lab and NVIDIA on AI tools deployed directly on on-premises Dell PowerEdge servers, which the speaker describes as easier, faster, cheaper, and better to deploy than in the cloud.
The video highlights several healthcare AI use cases, including Aires, an in-house GenAI radiology tool that helps radiologists focus on image interpretation and diagnosis. The speaker says some radiologists are seeing 40% efficiency gains. The video also describes work on GenAI and multimodal language models to predict a patient’s future, a capability the speaker says some call a digital twin.
Key takeaways
- AI can help healthcare move from reactive care toward earlier prediction and prevention.
- Northwestern Medicine is deploying AI tools on on-premises Dell PowerEdge servers with support from NVIDIA.
- GenAI radiology tools can help clinicians spend more time interpreting images and communicating diagnoses.
- Digital twin-style models can help hospitals predict patient outcomes and operational needs.
- The video frames the hospital AI factory as a way to make AI creation more accessible across healthcare systems.
AI use cases for proactive healthcare
| Use case | What the video says | Infrastructure relevance |
|---|---|---|
| GenAI radiology | Healthcare AI depends on sensitive clinical data that must remain protected. | Imaging AI needs accelerated compute, secure data access, and reliable deployment close to clinical workflows. |
| Proactive care | Imaging, prediction, and model evaluation can require significant processing power. | Predictive AI needs access to clinical data and infrastructure that can support model development and deployment. |
| Patient digital twin | Hospitals may need to keep critical workloads close to patient data and controlled environments. | Digital twin workloads need compute, storage, data integration, and governance to support prediction and simulation. |
| Hospital operations | Radiology and clinical AI workflows often involve large imaging and patient datasets. | Operational AI needs integration with hospital systems and infrastructure that can scale across use cases. |
| Hospital AI factory | Clinical AI requires oversight, access controls, and traceability. | AI factories need infrastructure, software, data, and services that allow hospitals to build and deploy AI repeatedly. |
FAQs
How is AI helping hospitals get ahead?
AI can help hospitals move from reactive care toward earlier prediction and prevention. In the video, Northwestern Medicine describes using AI to predict disease states months or years before they happen and to improve radiology workflows.
Why are hospitals running AI on-premises?
Hospitals may run AI on-premises when they need stronger control over clinical data, performance, deployment speed, and cost. In the video, Northwestern Medicine says its AI tools are deployed on on-premises Dell PowerEdge servers and describes that approach as easier, faster, cheaper, and better than deploying in the cloud.
How does Dell infrastructure support healthcare AI?
Dell infrastructure can support healthcare AI by providing the on-premises compute foundation needed to run AI tools close to clinical data and hospital workflows. The video specifically mentions AI tools deployed on Dell PowerEdge servers, supported by NVIDIA.
What is a healthcare digital twin?
A healthcare digital twin is a model that uses patient or operational data to help predict future outcomes. In the video, Northwestern Medicine describes using GenAI and multimodal language models to build a comprehensive predictor of a patient’s future, which the speaker says some call a digital twin.
Imagine a future where patients rarely end up at the hospital. This is not science fiction. Through the power of collaboration, AI is truly the key to unlocking the promise of technology in healthcare
AI gives us a tremendous opportunity to help us take better care of our patients. Healthcare can shift from being reactive to being fully proactive, predicting disease states months to years before they happen. Together with the Dell Innovation Lab and supported by NVIDIA, we're working on AI tools deployed directly on our on-prem PowerEdge servers, which we find much easier, faster, cheaper, and better to deploy than in the cloud.
Aires is a GenAI radiology tool developed in-house, which allows me to perform the most important and the most fun part of my job as a radiologist, which is looking at images, forming an opinion, and communicating a diagnosis. We're seeing some radiologists gaining 40% in terms of efficiency. The partnership with Dell Technologies and NVIDIA has been truly unique.
It became very clear that we held some critical knowledge that would unlock their ability to make better technology, and they held critical knowledge that would unlock our ability to better take care of our patients. Another specific project that we've been working on is using GenAI and multimodal language models to develop a comprehensive predictor of the future of the patient.
Some call this a digital twin as well. We can proactively predict what will happen and prevent these things from happening. Now we have a tool that the hospital can use to predict all aspects of its operations. This opens up completely new avenues. I mean, the possibilities are truly endless, and we're extremely excited about the future.
With Dell Technologies and supported by NVIDIA, we are democratizing access to the very tools that create AI. The term people like to use is AI factory. I want every hospital system worldwide to have its own AI factory, which ultimately is going to save thousands, if not millions, of lives.