Last month, dermatologists were told they had narrowly lost a competition.

“Man against machine,” a study by Holger Haenssle and colleagues, found that artificial intelligence known as deep learning convolutional neural network edged out 58 dermatologists in the photographic diagnosis of melanoma.

They were charged with differentiating melanoma from benign moles using images obtained via dermoscopy, a technique that allows dermatologists to view the skin through a high-quality magnifying lens with a powerful lighting system.

This story made headlines: “AI beats doctors at cancer diagnoses.” As dermatologists, we read the headlines with surprise and thought, “Aren’t we on the same team?”

To optimize the potential benefits of artificial intelligence while preserving the doctor-patient relationship, we believe that collaboration, not competition, is the winning strategy.

In 1960, the renowned American psychologist and computer scientist J.C.R. Licklider published his treatise: “Man-Computer Symbiosis,” which described how the two could have a mutually beneficial relationship.

In nature, he noted, the fig tree is pollinated by an insect that depends upon it for food. He then extrapolated the concept of symbiosis to computing. In his paradigm, humans “set the goals, formulate the hypotheses, determine the criteria, and perform the evaluations,” while computers “do the [routine] work that must be done to prepare the way for insights and decisions in technical and scientific thinking.”

Human-computer symbiosis provides a valuable lens through which to view the appropriate role of artificial intelligence in medicine. As AI becomes more powerful, rather than fostering competition, we should develop solutions that can be integrated into our practice.

The promise of artificial intelligence is considerable, with potential gains in efficiency, patient outcomes and access to care. But so, too, is the risk.

AI is only as “intelligent” as its inputs. The convolutional neural networks, as used here, are “trained” by viewing thousands of images, learning to recognize patterns alongside diagnoses and improving through feedback.

Inaccurate or incomplete training of this form of AI could result in misdiagnoses in patients.

For example, this study’s training set did not include most melanoma mimickers beyond moles. And it included only a small percentage of atypical moles, which are what dermatologists are commonly trying to differentiate from melanoma.

Moreover, a convolutional neural network trained to diagnose melanoma using images primarily obtained from light-skinned patients may disproportionately miss melanomas on patients with darker skin.

By working together, we can ensure that AI tools are valuable and safe for patients.

Only a human doctor can view a patient holistically — as a complex physical and emotional being. A patient is more than a single mole.

Evaluating the entire skin, a dermatologist searches for the mole that stands out — “the ugly duckling” — as a sign of melanoma.

In addition, the doctor and the patient are brought together by a goal not simply of diagnosing and managing a single spot or disease, but of healing. The word “sacred” is often invoked to describe the doctor-patient relationship.

As dermatologists, our shared humanity enables us to empathize with patients suffering from melanoma and other skin diseases.

Because the skin is a visible organ, the diseases we treat are often socially stigmatizing and have a profound impact on quality of life. Our training allows us to connect with patients in a unique way. For example, we may be the first to hold the hand of a patient with a severe, yet not contagious, autoimmune blistering disease resembling chickenpox, whom others are afraid to touch.

A valuable trajectory for artificial intelligence would be the development of tools that reduce our administrative and documentation burdens and afford us more time to communicate with our patients.

As doctors, we make a commitment not only to care about our patients as individuals, but also to care about the diseases that affect them. In dermatology, we are constantly working to deepen our understanding of skin diseases in the hopes of discovering new therapies.

While artificial intelligence may advance the current practice of medicine, it cannot identify, prioritize and solve future problems.

Before 1861, doctors pressed their ears against patients’ chests to listen to the heart and lungs. The stethoscope revolutionized medicine by providing doctors with auscultatory data on which to base diagnostic and management decisions.

As dermatologists, we applaud the work of Haenssle and other innovators to make strides toward early detection of melanoma. But words matter. We believe that rhetoric that sensationalizes a competition between humans and technology is counterproductive and has the potential to undermine the human doctor-patient relationship.

Throughout history, doctors have evolved in a symbiotic relationship with technology to confront the burden of human disease.

If artificial intelligence becomes another tool in our toolbox, it will support the time-honored lesson that man with machine is superior to either alone.

Carrie L. Kovarik is an associate professor of dermatology, dermatopathology and medicine at the University of Pennsylvania. Caroline A. Nelson and John S. Barbieri are chief residents in dermatology at the University of Pennsylvania Perelman School of Medicine.