A behind-the-scenes look at Midjourney’s medical scanner leaves many questions unanswered

Midjourney is stepping far beyond its roots in generative art with a bold new venture into medical technology. The company recently shared a behind-the-scenes look at their so-called medical scanner, a device they hope will revolutionize healthcare by providing cheap, detailed, and radiation-free imaging. This move signals a serious pivot for the AI giant, aiming to apply their computational prowess to physical hardware challenges.

According to reporting by The Verge, the device is essentially a large immersion tank filled with ultrasound gel. It features an elevator mechanism that lowers patients into the water while dozens of ultrasound probes surround them. The concept is intriguing on paper, offering a non-invasive way to capture comprehensive body scans without the high costs or risks associated with traditional MRI or CT machines.

However, the reality of the prototype is far less polished than the marketing suggests. Marcin Plaza, an engineer at Midjourney and the YouTuber who documented the tour, described the setup with unvarnished honesty. He noted that the system consists of ultrasound probes that were hacked apart and slapped onto what he called a glorified hot tub. This candid admission highlights the rough, experimental nature of the current stage of development.

The hardware stack is equally unconventional for a medical device. The system relies on off-the-shelf computers and Raspberry Pi units to process the data. While this approach keeps costs low and accelerates prototyping, it raises immediate concerns about reliability, regulatory compliance, and the robustness required for clinical environments. Medical devices typically undergo rigorous testing to ensure they can withstand continuous use and maintain precision under stress.

Perhaps the most pressing issue is the lack of proof that the technology actually works as intended. The video shows scans of imaging phantoms, which are standardized objects used to test equipment. While these tests validate that structures separate cleanly under controlled conditions, they do not demonstrate efficacy on human subjects. The gap between phantom validation and clinical utility is vast and difficult to bridge.

This project also underscores a broader trend in AI where companies are moving from software-only solutions to integrated hardware-software systems. Midjourney is betting that their ability to interpret complex visual data can compensate for the limitations of simpler hardware. If successful, this could democratize access to high-quality imaging, particularly in regions lacking expensive medical infrastructure.

Yet, the skepticism surrounding the project is understandable. The medical industry is highly regulated and slow to adopt new technologies without extensive clinical trials. Investors and healthcare providers will likely demand substantial evidence of safety and diagnostic accuracy before considering deployment in spas or clinics. The current prototype feels more like a proof of concept than a ready-to-use medical tool.

What this means for you

For professionals in healthcare or tech, this development serves as a reminder that AI hardware integration is still in its infancy. It is crucial to critically evaluate claims of disruption and look for peer-reviewed data rather than promotional videos. You can use this insight to stay ahead of the curve by monitoring how AI companies navigate regulatory hurdles.

Try this prompt with your AI assistant to analyze the regulatory landscape for AI-driven medical devices:

"Act as a regulatory consultant. Outline the key FDA or CE marking requirements for a Class II medical device that combines ultrasound hardware with AI image analysis. Highlight the specific challenges related to using consumer-grade components in a clinical setting."