Midjourney annonce un scanner ultrasonore corporel complet comparable à l'IRM, mais les experts en radiologie dénoncent : une exagération extrême, sans aucun fondement

IA image generation company Midjourney announces its entry into medical imaging, launching a device that immerses users in a water tank for a 60-second full-body ultrasound scan, claiming image quality "equivalent to MRI." However, five radiology and cardiology professors state that its technical claims are "extremely exaggerated" and "completely unfounded."
(Background: The secret society Dialog members list leaked, including Musk, Trump’s son-in-law, former Google CEO, Silicon Valley founder, U.S. Treasury Secretary)
(Additional context: Crazy? Mysterious whale actively burns 107 BTC to a black hole address, sparking three major community theories)

Table of Contents

Toggle

• 60-second water tank scan, ambition to rival MRI
• What happens when ultrasound hits bone and fat
• Experts: Likely more scam than transformation

A company with no prior experience in medical devices suddenly announces a "more powerful than MRI" full-body scanner, but five professors from Michigan, Washington, Wisconsin, and Jefferson University give nearly unanimous responses: no evidence, overhyped, far from mature.

60秒水槽掃描，對標 MRI 的野心

Midjourney CEO David Holz explained last week this business transformation: users step into a water tank, immersed for 60 seconds, with 40 Butterfly Ultrasound-on-Chip modules performing a full-body ultrasound scan, processed by AI and custom chips. The company describes this experience as "as relaxing as going to a spa," but claims the generated images are "as powerful as MRI," with Holz even hinting in an interview with The Verge that future systems might surpass MRI.

The plan is quite concrete: open the first spa in San Francisco by 2027, deploy 50,000 scanners by 2031, capable of 1 billion scans per month. Midjourney has invested $74 million, and signed a partnership agreement with ultrasound chip maker Butterfly Network in November 2025.

Medical director Tom Calloway told The Verge that the device is currently positioned as a "wellness product," not a medical diagnostic device. The company has confirmed this classification with the FDA, and cannot claim cancer screening or lifespan extension. The third-generation scanner is expected in 2028, claiming that image quality and speed will then be "worlds apart."

In its official blog, the company even cites statistics claiming "with enough early imaging screening, 30% of deaths and 50% of medical costs could be avoided globally." This announcement quickly sparked online debate, with many Silicon Valley tech commentators praising it as a "disruptive moonshot," a new way to "monitor body conditions continuously and for free." Most of these discussions come from outside the medical community.

超音波碰到骨頭和脂肪會發生什麼事

Ultrasound as a medical imaging tool has a fundamental limitation: it cannot penetrate bones and air-filled cavities. Simply put, sound waves reflect directly at interfaces with bone and air, preventing them from reaching deeper tissues. The thoracic, abdominal, and cranial cavities are all ultrasound blind spots. Additionally, fat tissue rapidly attenuates ultrasound signals, so image quality degrades significantly for larger users. The images shown by Midjourney so far are from slender subjects.

Although water immersion theoretically aids signal transmission, it requires extremely pure water free of bubbles, with specialized degassing devices. After each use, water must be replaced. Users also need to shave beforehand to prevent residual bubbles between skin and water from interfering with signals.

Scott Reeder, a radiology professor at Wisconsin, points out that current ultrasound techniques usually require over 30 minutes for regional scans, whereas MRI and CT, though longer, provide much higher quality information. He says achieving MRI or CT-level results "is a big leap," and current technology "is not yet mature."

Venkatesh Murthy, a preventive cardiology professor at Michigan, states that the resolution claimed by the company "is obviously theoretical," and the MRI equivalence claim "completely unfounded." He also notes a gap between regulation and marketing: "Most of the company's messaging isn't about body composition, but about cancer screening and lifespan extension." Since body composition measurement already has existing technology, he adds, "some scales are almost as accurate."

Experts: More likely a scam than a transformation

Among the five professors, the most severe words come from William Morrison, a radiology professor at Thomas Jefferson University. He describes the whole thing as a "vibe-based rollout," a strategy driven by emotion and expectation rather than actual technical data. He states that the current displayed images "are far behind existing CT and MRI," and that the water bath method has "almost been abandoned" in modern medical imaging. The whole thing "resembles more a marketing stunt than a real transformation," and directly says: "This might be more scam than innovation."

Matthew Davenport, a radiology professor at Michigan, also uses harsh words, calling the company's claims one of the "most exaggerated" he's seen. Mark Anastasio, an imaging scientist at Washington University, remains more reserved, believing that full-body ultrasound "is indeed feasible," but emphasizes "there's no evidence" that such ultrasound scans can match MRI.

Midjourney's fallback is regulatory positioning: since it is a wellness product, it is not subject to strict medical device regulations. But this also means users pay for scans that do not provide diagnostic-grade medical information. Reeder further warns that if people skip mammograms or colonoscopies because of this, "that would be concerning."

Davenport's criticism extends to ethics: "Launching unverified claims that are almost certainly unachievable is ethically problematic." He and Reeder co-authored an article in JAMA this year discussing the pros and cons of whole-body MRI screening, noting that large-scale imaging screening does not automatically lead to better medical decisions.