Photo By: National Cancer Institute

Artificial intelligence is becoming more common in healthcare. It can help organize information, summarize medical records and assist healthcare workers with certain tasks. The Military Health System is also looking at ways to use artificial intelligence in military medicine.

At the same time, there is a growing concern about the quality of information produced by AI. One term used to describe poor quality AI-generated content is “AI slop.” The term is often used for content that is produced quickly by artificial intelligence and may contain mistakes, misleading information or very little human review.

AI slop is often discussed in connection with social media, where large amounts of AI-generated pictures, videos and articles can appear every day. In healthcare, however, inaccurate AI-generated information could be much more serious because medical decisions depend on accurate information.

One area where AI is already being explored in military healthcare is medical documentation. The Defense Health Agency has been testing ambient technology that can listen to conversations between healthcare providers and patients and create a draft clinical note. The goal is to reduce some of the time healthcare workers spend documenting patient visits. 

This type of technology could be helpful. Doctors and other healthcare workers spend a significant amount of time documenting patient care. If AI can create a useful first draft, a healthcare worker can review it and make corrections instead of starting the note from the beginning.

The problem is that AI does not always get information right.

AI systems can sometimes create information that was not actually given to them. This is commonly called a “hallucination.” AI can also leave out information that was included in the original conversation or medical record.

A 2025 study published in npj Digital Medicine examined AI-generated clinical notes. The researchers found examples of both incorrect information and missing information. Some of these errors were considered serious enough that they could potentially affect patient care if they were not corrected. 

This is where the larger issue of AI slop becomes important. The problem is not simply that AI can make mistakes. People make mistakes too. The concern is that AI can produce large amounts of information very quickly, and that information can look professional even when parts of it are wrong.

Joanne M. Frederick, CEO of Government Market Strategies, points to a related problem in the way government organizations use technology. As Frederick puts it, “We shouldn’t use 21st-century technology to accelerate a 20th-century habit of layering new rules on top of old ones.”

Her point raises an important question for military healthcare. AI can make government processes faster, but speed alone does not solve an inefficient or complicated system. If an organization has too many overlapping rules, procedures and requirements, simply adding AI may cause the system to process those problems faster rather than actually fixing them.

This matters because military healthcare involves both medical care and government requirements. Healthcare workers have to work within established procedures while also protecting patient information and maintaining accurate medical records. Adding new AI systems to that environment creates another layer that needs to be understood and managed.

Frederick’s perspective also connects to the larger concern about AI slop. If organizations use AI to produce more documents, summaries, reports and information without improving the systems behind them, they could end up with more information but not necessarily better information.

The Military Health System has already taken steps to examine some of these risks. In 2024, the Department of Defense and the Military Health System conducted a “red team” exercise involving large language models. Participants looked for problems involving bias and vulnerabilities in AI systems being considered for military medical uses, including clinical note summarization and medical advisory chatbots.

This type of testing is important because AI systems need to be evaluated before people depend on them for important work.

AI could still provide real benefits to military healthcare. It could help healthcare workers organize information, reduce some administrative work and create drafts of medical documentation. These tools may allow healthcare professionals to spend more time focusing on patients.

However, AI should be treated as a tool that assists healthcare workers, not as a replacement for them. An AI system can create a draft, but a healthcare professional needs to check the information. If something is incorrect or missing, a person needs to be able to identify and correct the problem.

The future of AI in military healthcare will therefore depend on more than how advanced the technology becomes. It will also depend on how carefully organizations use it.

The goal should not be to use AI simply because it can make a process faster. The goal should be to determine whether the technology actually makes that process better.

That is the larger lesson behind AI slop. Technology can produce information faster than ever before, but producing more information does not automatically make the information more useful or more accurate.

For military healthcare, accuracy needs to come first. AI can help healthcare workers, but human judgment, review and responsibility still matter.

As Joanne M. Frederick’s observation suggests, the challenge is not simply bringing new technology into an old system. The real challenge is deciding which parts of the system should be improved before technology is used to make them faster.

AI may become an important part of military healthcare. The question is whether it will be used to create a better system or simply a faster version of the old one.