Last week, Midjourney, an AI startup best known for its image generator, made an unusual pivot: medical imaging.

The company announced a futuristic ultrasound scanner that would dunk users into a vat of water and, hopefully, produce “something as powerful as MRI” yet “as casual as a trip to the spa.” Midjourney says the goal is to help people live longer, better, and healthier lives. CEO David Holz has suggested the system could one day be better than MRI. Experts are skeptical. While several medical imaging specialists told The Verge they were not dismissive of the idea outright, they said Midjourney has shown little public evidence to substantiate its goals —especially for a technology that has been around for decades and has well-understood limits.

To call this a left-field move is an understatement. Midjourney is shifting from generating synthetic images online into the high-stakes, tightly regulated world of medicine — and definitely not-so-tightly regulated realm of wellness. It is not immediately clear how the venture relates to Midjourney’s existing AI business model, and AI is hardly mentioned in the blog post laying out the company’s plan. Midjourney’s head of medical Tom Calloway told The Verge the scanner uses AI and specialized chips to handle the “unthinkably huge amounts of data and processing power” required to “execute a scan.” Calloway said that AI is also used “to enable lossless compression and dramatically speed up processing.”

Online, the announcement was celebrated as exactly the kind of disruptive moonshot Silicon Valley loves, with some hailing it as the future of medicine and a path for cheap, constant monitoring that could save and prolong lives (while making Midjourney a boatload of cash). Much, though not all, of that enthusiasm came from outside medicine and radiology.

There is still “a long road ahead to generating high-quality images and then to understand the clinical value and demonstrate net benefit to patients.”

Radiologists, clinicians, and imaging experts who spoke to The Verge were less awestruck. Several said the concept is genuinely exciting and maybe even plausible. But they also said the idea is not as novel as Midjourney suggests, raised basic questions about its execution, and all asked the same crucial question of the company: Where is the proof?