The Rise of AI in Healthcare Shows Both Progress and Roadblocks

Artificial intelligence (AI) is increasingly shaping healthcare, from spotting diseases early to improving hospital operations. Managed Healthcare Executive reviewed articles covered on AI we’ve published from August to early November 2025 to see the main trends in the growing space. The findings show that while AI is promising, careful execution is needed to make it truly effective.

AI is being used to better understand diseases and improve patient care.

A study from Insilico Medicine, published in Aging, looked at aging and idiopathic pulmonary fibrosis (IPF), a serious lung disease that mostly affects older adults. Researchers of the study developed two AI tools: a proteomic aging clock and a model called ipf-P3GPT. The aging clock uses blood protein data to predict biological age.

This study found that patients with severe COVID-19, who are at higher risk for lung fibrosis, had a biological age about 2.77 years higher than healthy people. The ipf-P3GPT model analyzed gene activity in those with IPF and normal lung aging.

“Notably, more than half of these shared genes showed opposite regulation patterns: some were upregulated in IPF but downregulated during normal aging, and vice versa,” the authors wrote.

These findings suggest that IPF isn’t just accelerated aging but involves abnormal gene activity, offering new ways to develop treatments and biomarkers.

AI is also helping detect diseases earlier.

Shailja Shah, M.D., M.P.H., and her team created a natural language processing (NLP) tool that identifies precancerous conditions in the stomach and esophagus from electronic health records (EHRs). The algorithm was very accurate, between 97.5% and 100%, for detecting conditions such as Barrett’s esophagus and gastric intestinal metaplasia.

It also captured detailed information, such as the exact location and type of tissue changes, which can be missed in standard medical codes. This allows hospitals to identify high-risk patients for closer monitoring while avoiding unnecessary tests for low-risk patients.

However, AI is not free from problems. A study in npj Digital Medicine found that AI models in psychiatry often showed racial bias. The bias was strongest for patients with schizophrenia.

“Most LLMs exhibited some form of bias when dealing with African American patients, at times recommending dramatically different treatments for the same psychiatric illness and otherwise same patient,” Elias Aboujaoude, M.D., MA, from Cedars-Sinai, said.

This highlights the need to carefully test AI to make sure it does not worsen healthcare inequalities.

AI is also being used to improve hospital efficiency. Ambient listening tools can record patient visits and create clinical notes quickly, allowing doctors to spend more time with patients. Ben Scharfe of Altera Digital Health shared how AI can also create highly personalized patient education.

“Essentially, the way we can do this is we pull in official, vetted content … from sources that are curated by the care organization, and then we combine that contextually with the transcript of the encounter, as well as with relevant data from the patient’s chart,” Scharfe said.

This approach allows advice to fit each patient’s lifestyle, making it easier for them to follow medical guidance.

Even with these successes, experts urge caution.

Serge Perras, MBA, CIO at Abarca Health, said AI is still in a “hype phase.” He warned against overenthusiasm, like what happened during the dot-com boom and bust.

“Please look at it (AI) from a business perspective and a value proposition, not at what the technology is for the sake of technology, but at what the business problem is that you’re trying to solve,” he said.

At Abarca, an AI bot cut time for medication reviews from two to two and a half hours down to one, showing clear efficiency opportunities. However, Perras stressed that AI works best when applied to high-impact tasks.

Other challenges include data quality, regulatory issues and understanding how AI makes decisions.

“Nondeterministic means I don’t really understand how a decision was made,” Perras said, adding that many AI models lack transparency about what information influenced their outputs. He also said that pilot testing, collaboration across teams and staff training are key to using AI effectively.

Healthcare leaders such as Kedar Mate, M.D., co-founder of Qualified Health, stress intentional design.

“For me, this is all about the choices that we decide to make,” he said, sharing concerns about AI in healthcare that can be resolved with “deliberate, eyes-wide-open actions.”

“We have options around how to train those tools,” he added. “We can train them on biased information … or we can prune that knowledge base to look for highly reliable, valuable information that we believe is, in fact, factually accurate.”

From improving patient care to making hospitals run more smoothly, AI is no longer just a future concept—it’s here. However, its success depends on thoughtful use. Managed Healthcare Executive’s review shows that AI is best seen as a tool to support human decision-making, enhance efficiency and improve accuracy while ensuring fairness and equity in healthcare.