Michael Blackman, MD, MBA, says one of the risks of AI is that “It pulls information together that’s logical, presents it in a cohesive fashion — but… is it true?”

Blackman, now chief medical officer at Greenway Health, has spent years at the intersection of clinical practice and health IT. While AI technology is advancing at a near light-speed pace, his insights are grounded in what it all means for day-to-day operations. Blackman shared his thoughts about AI on a recent episode of the MGMA Insights Podcast with host Daniel Williams, senior editor at MGMA.

The Three Buckets: Will AI Help or Harm?

AI outputs may feel authoritative, but Blackman argues they fall into three distinct categories, only one of which requires real vigilance. “You end up with — especially from a decision support perspective — the AI falling into three buckets,” he explains. “Bucket one is, ‘Oh yeah, I recognize that. I just didn’t think of it.’” At the other end: “Bucket three is, ‘I recognize that’s wrong… that doesn’t make any sense.’ Easy to dismiss.”

The danger sits squarely in the middle. With Bucket two, Blackman says, “This is logical, but do you then look a level deeper to check... to make sure it’s really true?”

In other words, it’s not the obvious errors or the obvious insights that present risk — it’s the “convincing but unverified” middle ground. Leaders designing workflows need to identify where those gray-area outputs appear and ensure a review loop exists.

“I used one of the commercial AI tools… to help prepare for a paper,” he adds. “The five references it provided? None of them were real. Zero.”

AI Is Quietly Redefining Clinical Work

AI isn’t replacing clinicians, but it is redefining their role in subtle, operationally significant ways. Despite the need for caution, Blackman is unequivocal: AI is already creating measurable gains across clinical workflows —  especially with documentation.

With these ambient documentation tools now gaining traction, Blackman sees one of the clearest shifts in clinical workflows: “It has created a draft of the note,” Blackman said. “Does that mean you don’t have to read the note?… No, it doesn’t mean that at all.”

“You’re moving from being the author of the note to the editor of the note — at the end of the day, you still have to make a judgment, and then sign it,” he added. “It is much faster and… better interaction with the patient to build the note from the conversation automatically."

What’s more, AI may outperform human recall in real-time encounters. “It puts something in the note that [a clinician] didn’t hear,” Blackman notes. “They go back and look at the transcript… and sure enough, the patient said it.”

Beyond documentation, AI’s ability to surface and synthesize data is beginning to reshape chart review.
“You think about all the data that’s in a chart… tons and tons of information,” Blackman says. “Now you can query it… it can help surface information you may not [see].”

Task management is also evolving: “Helping organize… messaging back and forth to patients, pharmacies, interoffice,” he adds, pointing to early gains in prioritization and workflow efficiency.

Show Your Work

So, if the main risk is plausibility without accuracy, then the main skill should be thorough interrogation.

“Teachers used to tell us… in math class, show your work,” Blackman says. “Let’s ask the AI to show its work.”

This is a practical operational tactic — one that can be embedded into workflows and physician training.

“How did you get to that answer? What is that supporting evidence?” he continues. “And then more importantly, check to make sure the evidence is correct if you’re not certain.”

In practice, this could mean:

• Requesting citations or source reasoning from AI tools
• Verifying clinical suggestions against trusted databases
• Training clinicians to challenge outputs, not accept them

This aligns AI usage with existing clinical reasoning habits — differential diagnosis, evidence review, risk assessment — rather than treating AI outputs as endpoints.

AI as a Team Member with Different Communication Rules

Like any team member, AI requires adjusted communication patterns. “It absolutely becomes yet another member of the team,” Blackman asserted. "[But] if you want the information to be in your note, you have to vocalize it."

This represents a behavioral shift for clinicians accustomed to internal processing. “There are plenty of things that we used to do without vocalizing them,” he says. “Now you need to vocalize it.”

This affects both provider behavior and patient experience. Conversations become more explicit. Processes become more transparent. And teams must align on how information is captured and shared in real time.

Technology Overload Is Real — and Preventable

The rapid expansion of AI tools creates a new operational risk: cognitive overload. “What’s brand new this week is going to be old two months from now,” Blackman said. 

But the mistake, he argues, is adopting tools without anchoring them to clear problems. “You have to pick the problem you’re trying to solve,” he added. “It shouldn’t be the tech looking for a problem.”

Workflow integration is equally critical. “You want to make sure these tools are embedded appropriately in the workflow,” he explained, “so they’re not distracting… not causing extra cognitive burden.”

This is where many organizations fail — layering solutions onto already fragmented systems. Blackman points to a more sustainable approach: designing with AI from the outset.

“We’re trying to re-envision this from the ground up,” he says, “thinking about AI from the start so it’s in the right places in the workflow.”

Burnout Relief Comes in Inches, Not Miles

AI’s promise to reduce burnout is real but incremental. “There’s no silver bullet here,” Blackman emphasized. “It’s incremental pieces across the board.” Examples include:

• Documentation that “adds some time back”
• Systems that “tee up information for you”
• Tools that “appropriately prioritize incoming messages”

Equally important is the psychological shift in work. “It lets people focus on… the more cognitive challenging pieces… the more interesting pieces,” he said. 

That connects to a longstanding principle in healthcare workforce design: “Let people work to the top of their license; [AI] really helps enable that.”

Trust + Transparency = Better Patient Experience

AI’s impact ultimately lands in the exam room and in patient perception. “There are people who aren’t trustful of AI,” Blackman said. “They hear AI and they go, ‘Oh… I don’t want this involved in my care.’”

In the face of this resistance, transparency is the solution. “It’s incumbent upon us to talk with patients about what it is and what it isn’t,” he explained. When framed correctly, the benefits become clear to patients. “Tell a patient, ‘I’m going to use this tool, and it lets me focus my time on you,’” he added. “Most patients really like that.”

AI can also enhance clarity in care instructions. “Here’s your instruction summarized in a way they can really take advantage of,” he said, noting improvements in comprehension and compliance. And that combination — clarity, communication, focus — shapes whether AI strengthens or erodes trust.

Blackman’s invented exam answer earned half credit because it sounded convincing. Used thoughtfully, AI can do something far more valuable in healthcare: Give clinicians more time to listen to and connect with the patient in front of them.

Resources

• Connect with Dr. Blackman on LinkedIn.
• Greenway’s Novare platform (AI-driven EHR)
• AI by Design in Healthcare: Rethinking the EHR from the Ground Up, with Dr. Michael Blackman (MGMA Insights Podcast)
• "AI is slowly redesigning work in medical practices rather than replacing workers" (MGMA Stat Article)
• "Automatic for the people: AI moves for medical practices to boost the front office and access" (MGMA Stat Article)
• "AI governance in medical group practices: Rules for the humans in the loop" (MGMA Stat Article) 
  

Email us at dwilliams@mgma.com if you would like to appear on an episode. If you have a question about your practice that you would like us to answer, send an email to advisor@mgma.com. Don't forget to subscribe to our network wherever you get your podcasts!