Imagine a day when you wake up from a perfect six hours and thirty-six minutes of restorative sleep. Overnight, a wearable device or implanted chip has been continuously monitoring and capturing a comprehensive set of your biological and physiological variables. The ingested data from the variables collected is processed through an adaptive machine learning algorithm to create a physician-directed longevity score to increase health span—the length of time a person remains healthy and free from serious illness or disease throughout their life.

You receive an alert from your physician about new ST changes identified on the continuous ECG, with instructions to follow up within the next hour. Through ambient voice-activated technology, you call your physician, and she appears on the screen in front of you. However, she is not actually a live person but a digitally enhanced twin who is already acquainted with your entire health history.

You log in to your personalized HIPAA-certified patient portal wearing an Apple Vision Pro device, and within seconds, a voice announces, “The doctor’s digital twin will see you now.” An avatar of your physician’s digital twin appears in the virtual waiting room and welcomes you into a pleasant, relaxing consultation room where she provides you with the next steps regarding the ECG changes.

Although this scenario sounds more like a scene from a sci-fi movie, it is poised to become the new reality for the health care industry in just a few years, thanks to the rapid evolution of artificial intelligence (AI) and machine learning. Reflecting this shift, the American Medical Association (AMA) has embraced the term “augmented intelligence” instead of “artificial intelligence” to describe AI technologies. This choice underscores the AMA’s vision of AI as a tool designed to enhance and support the capabilities of health care professionals rather than replace them entirely.

The fast-paced expansion and adaptation of neural networks through AI will be the “clin-tech” that will change the delivery of health for millions of people, and the timing could not be more opportune. According to a report by the Association of American Medical Colleges (AAMC), we are on the brink of a dire physician shortage. By 2034, the projected shortfall ranges from 17,800 to 48,000 primary care physicians and between 21,000 and 77,100 specialists.

AI: a beacon of hope amid hesitancy

By automating routine tasks, AI can alleviate some of the pressures on overburdened physicians, allowing them to concentrate on more complex and critical aspects of patient care. Moreover, AI’s capability for precise, real-time monitoring and analysis can support early diagnosis and personalized treatment plans, further optimizing health care resources.

Despite the promising horizon that AI presents in addressing some of health care’s biggest challenges, its adoption has been slow among physicians. According to a recent survey by the AMA, although 77 percent of responding physicians believe AI will improve documentation for billing, medical charting, and visit notes, the actual integration into daily medical practice remains limited. Today, only one in five physicians leverages open-source generative AI such as ChatGPT in their practice.

Challenges and bias in AI integration

Skepticism towards AI integration in health care is healthy and warranted, given the myriad challenges it presents. Concerns range from data privacy breaches, misdiagnoses due to algorithmic errors, and the ethical implications of replacing human judgment with machine decision-making. Additionally, there’s apprehension about the lack of clear regulatory frameworks and standards for AI in health care, which could lead to inconsistencies in patient care.

Another issue to overcome is detecting and avoiding bias within AI. Algorithms depend on data for their training, and if the data used for training is biased, the algorithms may inadvertently perpetuate and magnify those biases. One example of this is evident in a 2019 study where a widely used AI system for selecting health insurance patients for extra resources required African-American patients to be significantly more ill to access the same resources as their Caucasian counterparts. Similarly, diagnostic bias impacts the reliability of AI in medical assessments. One of the most well-known examples relates to facial recognition algorithms utilized in dermatology that demonstrate lower accuracy rates for individuals with darker skin tones.

Creating responsible AI in health care

To effectively counteract these issues and ensure that AI doesn’t further entrench bias within health care, the teams responsible for creating AI health care technology should be as diverse as possible. This diversity allows for a comprehensive examination of potential biases from various perspectives and experiences. AI algorithms must also remain transparent, allowing for scrutiny and accountability of any biases or errors, with rigorous testing and monitoring in place.

More broadly, clear regulatory guidelines can help standardize the development, testing, and deployment of AI, which will assist in building trust among health care professionals and patients alike. Lastly, AI integration must enhance and improve the health care delivery process rather than burden it. Given that clinicians already navigate significant time constraints and extensive charting responsibilities, it’s crucial that new technologies do not add to this existing workload.

Caution is both appropriate and necessary. However, considering AI’s profound capacity to enhance the work of health care professionals and improve patient outcomes, we bear a responsibility to approach this technological frontier with curiosity, open-mindedness, and a readiness to embrace change. Once the proper safeguards are in place, we will usher in a new era of health care that will redefine the limits of medical care and patient well-being.

Scott Ellner has been a general surgeon for over 20 years, and can be reached at PEAK Health. He has transitioned into health care executive roles due to his passion for patient safety, quality, and value-based care delivery. His authentic leadership style inspires team members to navigate challenging situations, such as resistance to change and innovation, in order to bring about meaningful transformation. Most recently, he served as the CEO of Billings Clinic, the largest health system in Montana. During his tenure, Forbes recognized the clinic as the best place to work in the state. It was also at that time that he formulated a strategic growth plan that included the development of a level 1 trauma network and a rural-based clinically integrated network.