Artificial Intelligence is becoming more than just a trend. Soon, it seems that A.I. features will be available in, and integrated into, every part of our lives.

One field currently interested in implementing A.I. into its everyday workflow is healthcare. Advocates for this shift say that using artificial intelligence can alleviate some of the most burdensome aspects of medical practice, while those against the idea claim that there are manifold issues that could come with a wider implementation of A.I. systems in the healthcare industry.

The situation is certainly nuanced, and in this article, we’re going to go over what advocates are claiming about A.I., the reality of using A.I. in healthcare, and the possibility for a middle ground to emerge between A.I. users in the medical field and those skeptical of the technology.

The Promises of A.I. in Clinical Documentation

One of the benefits of A.I., advocates say, is its potential to streamline and automate repetitive tasks. If properly implemented, this could mean improved efficiency for workers in the healthcare industry by allowing them to simplify the documentation process.

There is some evidence to support this. Research published in September 2024 by JAMA Network Open found that “approximately half of clinicians using the AI-powered clinical documentation tool based on interest reported a positive outcome, potentially reducing burnout.”

Next, there are some who say that implementing A.I. systems can increase accuracy. This is because A.I. programs are generally good at processing large datasets quickly, allowing them to cross-reference information, minimize discrepancies, and, hopefully, enhance the overall quality of medical records.

Third, some say that the naturalistic tone and ability of A.I. to detect nuance more than a normal chatbot or call tree could allow for improved patient experiences. By providing patients with A.I. designed to be engaging and empathetic, they may be able to better convey the information they need to provide prior to a visit. Furthermore, once a patient is in the office, A.I. transcription and summarization software could create a better record of a patient’s issues than a doctor taking notes.

Finally, advocates argue that A.I. systems are scalable and adaptable, meaning they can be customized to meet the needs of different specialties and practices. From primary care to surgery, experts say that these systems have the potential to be tailored to capture the nuances of various medical fields.

The Reality: Challenges and Limitations

While it’s easy to look at the above through rose-colored glasses, it’s important to note that many of these benefits are, as of the time of writing, theoretical, and many have reported mixed experiences actually using A.I. in their medical practice.

For example, many have reported that their efficiency did not improve using artificial intelligence. For example, a recent study published in the New England Journal of Medicine noted that “our findings suggest that the tool did not make clinicians as a group more efficient.” In fact, for certain users, adapting to the technology could add to their workload rather than reduce it, as implementing these systems takes time and effort.

Similarly, there may be usability issues that come with implementing A.I. Not all clinicians are comfortable with technology, and for those less tech-savvy, navigating AI-driven documentation systems can be a source of frustration. Plus, as the technology is still in its relative infancy,  Learning curves and technical glitches can further diminish the tools’ utility in assisting physicians.

Another notable concern is the potential for there to be a loss of nuance in documentation — or worse, A.I.’s tendency to hallucinate could cause confusion or issues providing care. In October 2024, the Associated Press published a story noting how artificial intelligence-powered transcription programs had a tendency to add incorrect details about patients, which could present issues in providing care.

On that topic, there are lingering ethical and legal concerns about using artificial intelligence. If an A.I. system introduces an error into a patient’s record, determining responsibility can become a complex issue. Furthermore, there are concerns about patient privacy and data security, especially as sensitive information is processed by A.I. algorithms.

Finally, it should be noted that implementing A.I. tools can be costly. These costs are not only financial, but also include the time and effort required for training and system integration. This can create barriers for smaller practices or resource-limited health systems.

What’s the Middle Ground?

The disparity between the promise and reality of A.I. in clinical documentation underscores just how important it is to take a balanced approach before implementing any A.I. system.

Before an A.I. system is put into place, staff must receive customized training and support for whatever system they are trying to implement. Furthermore, one should always look at the A.I. tools as a collaborative partner, not a replacement for actual care; for example, an A.I. could handle routine documentation, while physicians could retain control over critical and nuanced aspects of patient care.

Finally, emphasis should be put on oversight of any A.I. program, not only to mitigate ethical and legal concerns, but to allow the people actually using the programs to provide feedback in order to determine the strengths and weaknesses of the programs.

Looking Ahead

A.I. has the potential to revolutionize healthcare documentation. That said, there is still considerable work that needs to be done before these tools can, and should, be fully implemented into any practice. By leveraging the lessons learned from early adopters and continuously monitoring and refining these tools, the industry can—hopefully—use these tools to move closer to achieving an ideal balance of efficiency, accuracy, and patient-centered care.