Midjourney's Body Scanner Pivot Raises Red Flags

Leading experts are demanding substantiation for the ambitious claims made by the AI startup, Midjourney, as it ventures into new territory.

Known primarily for its generative AI image tools, Midjourney recently announced an unexpected pivot into medical imaging. The company unveiled a visionary ultrasound scanner designed to immerse users in a water vat, promising results "as powerful as MRI" yet "as casual as a trip to the spa." Midjourney states its overarching goal is to enhance longevity, well-being, and health. CEO David Holz has even posited that the system could eventually surpass MRI capabilities. However, this bold vision has been met with considerable skepticism. While some medical imaging specialists acknowledge the inherent potential of the concept, they emphasize Midjourney's notable lack of public evidence to support its claims, particularly for a technology with well-established principles and limitations that has existed for decades.

This strategic shift represents an extraordinary departure for Midjourney, moving from the realm of synthetic online image creation into the highly regulated and high-stakes world of medicine, and the less stringently controlled sphere of wellness. The connection between this new endeavor and Midjourney's existing AI business model remains somewhat opaque, with artificial intelligence receiving minimal mention in the company's initial blog post detailing the plan. However, Tom Calloway, Midjourney's head of medical, clarified to The Verge that the scanner indeed leverages AI and specialized chips to process the "unthinkably huge amounts of data and processing power" required for scans, also employing AI for "lossless compression and dramatically speed up processing."

The announcement ignited fervent enthusiasm online, hailed by many as precisely the kind of disruptive "moonshot" that Silicon Valley champions. Proponents celebrated it as the future of medicine, offering a pathway to affordable, continuous monitoring capable of saving and extending lives, while also positioning Midjourney for substantial financial gain. Much of this excitement, however, originated from outside the established medical and radiology communities.

In contrast, radiologists, clinicians, and imaging experts interviewed by The Verge expressed a more reserved reaction. While several found the core concept genuinely exciting and potentially plausible, they underscored that the underlying idea is not as novel as Midjourney suggests. They also raised fundamental questions regarding its practical execution, culminating in a unified demand for the company: "Where is the proof?"

This critical question permeates every facet of Midjourney's ambitious vision, from its direct comparisons to MRI, to the projected volume of scans, the scanner's design, the fidelity of its images, and the optimistic hope that frequent imaging could significantly improve health and prolong lives.

At the heart of the "Midjourney Medical" offering is an advanced ultrasound scanner that submerges users, standing on a platform, into a water vat. Once immersed, a ring of underwater sensors and scanners emits sound waves into the body, capturing the returning echoes to generate internal images. The company draws an analogy to dolphin echolocation, aiming for a scan duration of "no more than 60 seconds." This contrasts sharply with typical ultrasound procedures, which can take 30 minutes or more depending on the specific test and body part. MRI scans, which utilize powerful magnets and radio waves for detailed internal imagery, often require even longer and can be uncomfortable, necessitating patients to remain motionless within a confined, noisy tube. Other detailed imaging methods like CT scans, which employ ionizing radiation, pose safety concerns for unnecessary or repeated exposure.

Midjourney asserts that its scanner is designed to empower individuals with more data about their bodies, facilitating improved health decisions. The company draws parallels between its process and images and those produced by MRI, even citing statistics that claim "the world could avoid 30% of all deaths and 50% of all healthcare costs" if early imaging were more widely available.

Despite this medical-centric rhetoric, Midjourney is not initially positioning the scanner as a diagnostic medical device, citing the extensive FDA clearance processes and clinical trials required for such applications. (The company does, however, indicate plans to expand into medical uses in the future.) Instead, it frames the device as a tool for personal information, intending to integrate the machines into spas where scans would become "a side-effect" of visits. Concept art depicts opulent golden rooms and pools, a stark contrast to traditional clinical environments.

"The technology is super cool," remarked Venkatesh Murthy, a professor of preventive cardiology, internal medicine, and radiology at the University of Michigan Medical School. He conveyed his excitement about the development of working prototypes for such technology, acknowledging that the underlying concept has existed for a considerable time. However, Murthy cautioned that "a long road ahead to generating high-quality images and then to understand the clinical value and demonstrate net benefit to patients" remains. He stated that many of Midjourney's "claims about resolution are clearly theoretical" and its suggestions of "MRI equivalent are completely unsupported," noting that the images presented by the company thus far "are decidedly low-res." (Midjourney's blog post, however, claims, "All of these images come together to cover a 3D map of your body, down to a fraction of a millimeter, that looks a lot like today’s MRIs.")

Mark Anastasio, a professor of imaging sciences at Washington University in St. Louis, characterized Midjourney's foray into medical imaging as an "exciting development," deeming the ultrasound scanner concept plausible. He highlighted ultrasound's extensive history as a diagnostic tool and recent progress in whole-body prototype systems using similar methodologies. Nevertheless, Anastasio stressed the imperative for validation before making any medical or diagnostic claims, adding that "there is no current evidence" to suggest such detailed ultrasound scans could rival MRI. Matthew Davenport, a professor of radiology at the University of Michigan Medical School, offered a more direct assessment. He found Midjourney's published images "interesting" and acknowledged a potential market for body imaging. Yet, he told The Verge that the company's "claims are wildly unsubstantiated," describing them as "perhaps the most grandiose" he has encountered.

Addressing the comparisons to CT and MRI scans, Calloway clarified to The Verge: "Every modality has its strengths and weaknesses. MRIs are powerful but expensive and slow. CT scanners are excellent for things like bone and lung imaging but expose patients to ionizing radiation. The first-generation scanner is fast, affordable, radiation-free, and great at imaging soft tissues like muscle and fat for body composition — but it won’t be a replacement for either of these technologies."

Scott Reeder, a professor of radiology at the University of Wisconsin School of Medicine and Public Health, also found the concept "innovative" but underscored that it "is not a proven technology," pointing to a lack of evidence on Midjourney's website to support clinical utility. He deemed it "a stretch" to compare these ultrasound scans to MRI or CT, which yield distinct types of information and possess different limitations. Reeder explained that ultrasound relies on sound waves traversing the body, making air and bone significant impediments. This could hinder imaging of body parts obstructed by or containing these elements, such as the gut and pelvic organs, especially without a technician to reposition probes. Imaging the head would also prove challenging, though Midjourney's scanner does not fully submerge users.

William Morrison, a professor of radiology at Thomas Jefferson University, expressed the deepest skepticism regarding Midjourney's proposal. He dismissed the entire initiative as a "vibe-based rollout," supported by images that "are markedly limited compared to existing technology such as CT and MRI." Morrison characterized ultrasound as a "very old technology" and noted that the water bath approach had largely been abandoned due to the physical limitations of sound waves, making its implementation difficult.

The water itself presents another challenge. Morrison explained that it would need to be "completely pure, without bubbles or dirt" to ensure high-quality images, an expensive and potentially customer-specific maintenance requirement. Dirt and hair could also trap air or deflect sound waves, Morrison noted, implying that optimal imaging might necessitate customers being "scrubbed and completely shaven." (In response to concerns about water purity and bubbles, Calloway stated to The Verge that "the bubbles are removed from the water with a large degasser located under the tank.")

Morrison further echoed Reeder's fundamental physics concerns: sound waves struggle to penetrate deeply enough and can be obstructed or distorted by air and bone. He added that fat can rapidly attenuate ultrasound waves, requiring signals to travel through more tissue, thus raising questions about the tool's effectiveness for individuals with larger body types. "I noticed that the images provided by the company are acquired on very thin people," he observed, suggesting a deliberate choice that implies lower image quality for larger patients. Calloway acknowledged this as a challenge also faced by CT and MRI scans, adding that Midjourney is "actively working with our partners" to address the issue.

Consistent with the marketing strategies of many AI health products, Midjourney has strategically positioned itself to circumvent numerous criticisms. By launching the scanner as a wellness product rather than a diagnostic medical device, and by initially offering only "detailed body composition maps," Midjourney subtly implies future equivalence with MRI images, extended lifespans, and improved health, without immediate regulatory pressure to provide definitive proof. The company states its intention to submit "regular test results" to the FDA to progressively expand authorized use cases.

Murthy concurred that "body composition is plausible" as a primary application, but highlighted that "much of the messaging [from Midjourney] isn’t about body composition but about cancer screening and overall lifespan prolongation." He pointed out that body composition can already be measured with existing technologies, noting that some weighing scales "will give you similar measurements with slightly less accuracy."

Uncertainty also surrounds how Midjourney will articulate this distinction to customers, and how clearly it will differentiate between body composition data and medical information. Morrison anticipated that the company would likely implement disclaimers to mitigate liability should the scanner fail to detect something critical. Midjourney refrained from providing specific details, though Calloway assured that the company is "being highly intentional" about the scanner's user interface.

"We’re strictly bounding the system to general wellness, building beautiful, detailed 3D maps of body composition," Calloway affirmed, adding that Midjourney has "confirmed this classification with the FDA and are operating squarely in that lane." He stated that future capabilities would be introduced in collaboration with regulators and medical experts. Midjourney's blog already outlines multiple planned generations of the scanner, with a third generation slated for 2028, promising that the technology will become "serious" and "image quality and scan times will be night-and-day." By 2031, Midjourney aims to deploy over 50,000 scanners globally, sufficient, it claims, to provide monthly scans to a billion people. The company anticipates opening its inaugural spa location in the "heart of San Francisco" in 2027.

Reeder expressed concern that "patients replaced proven screening technologies with this technology," citing hypothetical scenarios where individuals might forgo essential screenings like mammograms or colonoscopies, believing a Midjourney scan provided adequate assessment. Furthermore, even if the technology eventually proves capable of disease detection to warrant medical marketing, widespread mass scanning is not automatically a beneficial approach. The topic of elective full-body imaging is already contentious, a debate both Reeder and Davenport explored in a recent Journal of the American Medical Association article on MRI screening.

Ultimately, the core apprehension surrounding Midjourney's entry into medicine revolves around the imperative for proof. While the aspiration to develop technology that enhances human health is commendable, it underscores the critical importance of verifiable evidence. Noble intentions alone fueled the captivating vision of companies like Theranos. It is precisely why companies entering this field must possess robust evidence before proclaiming a transformation so profound it borders on the implausible.

"The interest in improving health through imaging is wonderful," Davenport stated. "The race to market with unproven claims that almost certainly will not prove true is ethically problematic." Morrison's assessment was even more severe: the entire initiative, he contended, feels like an advertising campaign. "It makes me think that this may be more of a grift than a pivot."