- NASA’s AI medical tool works where Earth-based doctors simply can’t reach
- Deep space has no signal, that’s why NASA built its own offline doctor
- RamaLama runs AI models the same way containers run software, unsurprisingly
Astronauts aboard the International Space Station (ISS) currently rely heavily on ground-based doctors when medical problems arise hundreds of kilometers high.
That arrangement works reasonably well in low-Earth orbit, where communications delays remain short enough to allow for near real-time consultation sessions.
It becomes much less practical once crews travel beyond Earth orbit, where signals can take minutes rather than seconds to arrive.
An AI doctor created to work without an Internet connection
Researchers at NASA’s Johnson Space Center in Houston are now testing a clinical decision support system called the Crew Medical Officer Digital Assistant, or CMO-DA.
The system is designed to help astronauts diagnose and treat medical conditions during deep space missions, where real-time communication with doctors on Earth may be limited or completely impossible.
Its engine is RamaLama, an open source tool backed by Red Hat created to simplify the way developers run and serve AI models on various hardware environments.
According to Red Hat, RamaLama treats AI models as container images, running them in isolated, security-first environments using Open Container Initiative-compliant containers that are portable and predictable across hardware.
This approach allows CMO-DA to perform what the team calls multimodal inference: processing both large language models for complex medical reasoning and vision language models for image-based symptom analysis.
Therefore, the system can evaluate both written descriptions of symptoms and visual data without requiring a connection to a terrestrial cloud server.
That offline capability is not a convenience feature but a mission-critical requirement, as communications delays in deep space make reliance on the cloud truly dangerous for crew health outcomes.
Tests are currently running on HPE hardware, specifically the Earth-based twin of the space computer already aboard the ISS, providing researchers with a reliable Earth-based replica of the actual deployment environment.
Using open source tools, NASA researchers have built a reproducible and auditable system, which the team describes as essential for human safety in mission-critical environments.
From ground testing to the ISS and beyond
Once the ground testing phase is complete, the CMO-DA will be demonstrated to NASA leadership to evaluate its subsequent deployment aboard the International Space Station.
The next version of the system will integrate Red Hat Enterprise Linux AI, known as RHEL AI.
This aims to provide a stable and solid foundation for scaling and managing containerized AI applications in remote and extreme environments.
RamaLama itself was built with the stated goal of making AI “boring,” that is, reliable, predictable, and unglamorous in the best possible sense for mission-critical applications.
The same architecture being tested for astronaut health could eventually serve as a model for providing medical assistance in the most remote areas of Earth.
It’s still unknown if the CMO-DA will eventually evolve into something like Star Trek’s portable Tricorder.
What is known is that an open source artificial intelligence tool is already diagnosing symptoms aboard a hardware replica currently orbiting Earth.
Follow TechRadar on Google News and add us as a preferred source to receive news, reviews and opinions from our experts in your feeds. Be sure to click the Follow button!
And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form and receive regular updates from us on WhatsApp also.
