AuntMinnie.com · The PACSMan Pontificates, Episode 3 - How real is AI?
The English words "The pen is mightier than the sword" were first written by novelist and playwright Edward Bulwer-Lytton in 1839, in his historical play Cardinal Richelieu. Jamaican political leader and orator Marcus Garvey expanded on his statement by saying “The pen is mightier than the sword, but the tongue is mightier than them both put together.”
If we really want to get technical, the Assyrian sage Ahiqar, who reputedly lived during the early 7th century BCE, coined the first known version of this phrase. In the Teachings of Ahiqar, dating to about 500 BCE, Ahiqar stated, "The word is mightier than the sword.” Greek playwright Euripides, who died c. 406 BCE, is supposed to have also written a version of this: “The tongue is mightier than the blade." Lastly, the Islamic prophet Muhammad is supposed to have said "The ink of the scholar is holier than the blood of the martyr."
Now, I’m not big on shedding my blood or being a martyr for PACS and AI any time soon, so before I even consider writing about AI, I usually put on a flak jacket and Kevlar helmet, have a commando knife and AR-15 with four spare clips, an emergency locator beacon for my editor to find me about questions on what I wrote, and some rosary beads as a nod to my Catholic upbringing. After all, you can never be too careful.
Few topics are as divisive as AI. Those on both sides of the coin are pitbulls defending their positions to the death. KLAS has said that over half of clinical-care organizations have adopted ambient AI tools to draft visit notes and transcribe patient interactions, drastically reducing administrative burdens. Repeated daily use of AI by physicians has nearly tripled year-over-year.
Physicians and nurses frequently use generative AI to instantly summarize medical literature, analyze patient data, and draft patient education materials. And AI-driven platforms direct patients to the most appropriate care settings, predict bed availability, and manage multiagent systems that handle end-to-end hospital processes. And what about imaging and diagnostics? We are on the board as well, but adoption is a fraction of what is found in the other areas.
As health systems expand their use of generative AI, IT leaders face a defining challenge: how to establish governance over AI-related data flows without stifling the clinical innovation these tools can deliver. Many organizations are still working to gain visibility into how staff interact with AI applications -- including unsanctioned tools that may expose protected health information.
A collaborative approach
The path forward requires more than policy; it demands a collaborative approach in which IT and security leaders partner with clinicians to create frameworks that enable responsible use rather than drive adoption underground. There is also a balancing act between moving quickly on AI and maintaining rigorous data protection. Adopting change management strategies that help build trust across clinical and technical teams is crucial in evolving from a mere gatekeeper to enabler in an era of rapid AI adoption.
Few would debate that AI is moving faster than most health systems can keep up with. But the real challenge isn't adoption. It's readiness.
Similar to what KLAS found, HIMSS indicated that 86% of organizations in the US are deploying or planning AI-enabled clinical decision support. Interestingly, only 7% have a mature governance strategy in place. At the same time, just 21% of leaders believe their data is being fully leveraged. The gap is clear.
AI depends on strong data foundations, governance, and a workforce prepared to use it effectively. Without that, even the most advanced technologies struggle to scale or deliver meaningful impact.
Today’s AI revolution has often been compared to the Industrial Revolution, arguing that workers, human dignity, and the common good must remain central as technology reshapes the economy. Has it? That depends on who you talk to and their perspective. A lot of it is how, or in this case if, governance is done. And governance within medical imaging? Certainly talked about extensively, but still needing more extensive and expansive real world development.
With medical imaging AI, there are some amazing algorithms available and more than a few dogs looking for adoption. Those easiest to justify are ones that impact results reporting; the hardest, those that deal with diagnosing some obscure Bora Bora disease for which they want $1,500 per use for the algorithm to detect. Most fall in between.
Sad reality
The sad reality is that most AI companies have limited to no marketing budgets. The only way you find out about them is a press release in some obscure journal who picks up on it that few read, a post on a website that most ignore, a LinkedIn post, a storyboard at a trade show, or other ways you can communicate on a zero or near-zero budget.
Now, this isn’t the entire market by any means, but if you can name more than 30 of the 800+ FDA cleared medical imaging companies seen on the AI-enabled medical devices list you earn a prize. Most can name those who have big marketing budgets and have sites where their product is implemented, are in the news often, or have high profile people promoting their products … but what about the rest? They are out there plodding along.
This also doesn’t include the plethora of homegrown AI algorithms popping up here and there that aren’t offered commercially and as such don’t require anyone's blessing. When I look at the lists I have to ask how many lung cancer or mammo screening AI products are needed? 10? 20? More? And what makes one better than the other besides the training done on the AI model and feature set? Is one worth 10 or 20 times what the other charges even with added features?
Every day you read about a new company or something new in the market. Most are blessed by some of the best-known AI evangelists. Seriously, though, have you ever heard of one of these people spreading anything but the gospel of prosperity with AI? Or those sowing seed faith where if you invest $x you will get 100 times the reward -- in LLM’s name. We have so many Joel Osteens, Tim and Tammy Faye Bakkers, Kenneth Copelands, Benny Hinns (a favorite for this Orlando boy), and so many others preaching the gospel of AI and how AI can and will change your life and practice.
Have any of them ever said a bad word about AI among them? They treat AI doubt as the Bible treats doubt -- not as an unforgivable sin, but as a normal human struggle and a step toward deeper faith. Even the Pope is getting on the AI bus. I actually like what he has to say.
Pope Leo XIV released his first major teaching document focused on AI, warning that AI development is moving too fast without enough ethical oversight. In the encyclical Magnifica Humanitas, he argued that AI should serve humanity rather than profit, power, or warfare, while pushing for stronger regulations and accountability from both governments and tech companies. The encyclical is already being viewed as one of the strongest global statements on AI ethics so far. Sing it, Pope Leo.
If I had read about this a bit sooner, I could have requested an audience with His Holiness over a plate of pasta, since I was in Rome just up the street from the Vatican just a few weeks back on vacation. We visited the real Italy -- Sicily -- where the Catholic influence can not be denied. Nearly 300 churches in Palermo alone to serve over 4 million Catholics.
Rather than rejecting the technology itself, Pope Leo XIV is pushing for a future where AI remains aligned with human values, transparency, and accountability instead of becoming another unchecked race for dominance. I love the fact that the Pope criticized the growing concentration of power among a small number of AI companies and warned against allowing private firms to define morality or ethics on their own. That said, he stands a better chance of … well, I’m not gonna go there. My space in hell is already reserved and I don’t need to get any closer to the fire by taking on the pontiff.
Real value
Even though there are over 800 vendors in the medical imaging AI market the question is no longer whether to invest in AI. It’s how to do it in a way that delivers real value for patients, clinicians, and health systems. Those that have the most compelling stories and can show sites where patients, clinicians, and health systems can see positive results from using AI will no doubt get the lion’s share of the investment dollars out there.
How much that is remains to be seen, though. Even though there is a lot of money floating around out there, investors are leery of companies no one has heard of, let alone products that may or may not show a return on investment within a reasonable time frame. What is reasonable remains to be seen. Five years is pretty much the de facto standard. Once again, while over eight of our 10 facilities in the U.S. are deploying AI in the form of AI-enabled clinical decision support, just one in five leaders believe their data is being fully leveraged.
With medical imaging, the total percentage of those using AI, at least in the U.S., is significantly less than 10%, with the vast majority of those either academic facilities or those who use devices with AI embedded in them. I agree with the comments made at a recent seminar put on by the Association for the Advancement of Medical Instrumentation where panelists addressed AI. While most of it was directed at extending imaging hardware usage, the comments were very relevant to medical imaging software as well.
“We’re standing at a fascinating time,” said moderator Robert Kerwin, general counsel for the International Association of Medical Equipment Remarketers and Services. “On one hand, AI promises to revolutionize patient care. On the other hand, there are real challenges with financial and cybersecurity concerns.”
Those challenges are already affecting hospitals’ ability to invest in new technology, said panelist Mike Powers, network director of biomedical engineering at St. Luke’s University Health Network in Bethlehem, PA. He pointed to ongoing cuts in Medicare and Medicaid reimbursement, noting that many health systems continue to serve patient populations heavily dependent on those programs.
“We’re struggling with keeping the door open,” Powers said. “We want to invest in new technologies and AI will supposedly help us save money. But how do we get from here to there?” As a result, many hospitals are extending the life of existing equipment rather than replacing systems outright. Panelist Stephen Grimes, managing partner and principal consultant at Strategic Healthcare Technologies Associates, described growing interest in “AI wrappers,” which are software layers designed to enhance legacy systems without requiring full replacement.
Now, this isn’t the case with all facilities everywhere but certainly many -- I’d say the vast majority -- can relate to the comments made. People are using medical imaging AI successfully but after a decade since it was introduced, the adoption pace is much slower than anticipated. How the end users or the facility are also justifying and paying for this remains to be seen, too.
You have some areas like mammo where AI is offered as a $40 service “upcharge” to a screening or diagnostic study. That’s fraught with challenges that will be addressed in a future podcast.
Stay tuned. With AI it only keeps getting better.
Michael J. Cannavo is known industry-wide as the PACSMan. After several decades as an independent PACS consultant, he worked as both a strategic accounts manager and solutions architect with two major PACS vendors. He has now made it back safely from the dark side and is sharing his observations.
His healthcare consulting services for end users include PACS optimization services, system upgrade and proposal reviews, contract reviews, and other areas. The PACSMan is also working with imaging and IT vendors to develop market-focused messaging as well as sales training programs. He can be reached at [email protected] or by phone at 407-359-0191.
The comments and observations expressed are those of the author and do not necessarily reflect the opinions of AuntMinnie.com.
