The rapid progression of artificial intelligence in the past decade, and particularly in the past year, has sparked many of the same questions in doctors and patients alike: what role will this technology play in the medical world? Is it reliable enough to be trusted? How will the shift toward AI affect the role of doctors, and to what effect?

To understand the developing role of AI in the medical field, you must first understand the other ways in which the medical landscape is evolving. The Internet age has led to an exponential increase in shared medical knowledge—in 1950, it took 25 years for our knowledge to double. Today, it only takes 73 days. In order to keep up with the amount of relevant medical news and data that is published every day, students would need to study quite literally over 24 hours a day.

As a result of this knowledge boom, medical students are increasingly drawn to becoming specialists. In the 1950s, around 1 in 20 residents decided to specialize. Today, that number is closer to 17 out of 20.

The purpose of introducing AI to hospitals stems from its ability to rapidly sift through big, complex data and draw meaningful conclusions. If the AI understands how to approach learning about new topics, and is well-trained enough to do so reliably and accurately, then medical students and professionals can begin to source more knowledge from AI synthesis and analysis. This will allow them to get a more comprehensive exposure to relevant medical news as it develops.

Artificial intelligence will also play a significant role in the future of diagnosis. The ability to compare a patient’s symptoms, history, and genetics against thousands of other patients can provide diagnoses and potential treatments that a doctor can’t see. AI could also identify potential problems, reactions, and other trends. For example, AI can pinpoint why a treatment is or isn’t effective for certain patients. This can be vital information for those who urgently need care.

Naturally, there is still skepticism towards the idea of artificial intelligence in medicine. It’s not unheard of for AI creations to learn bias or misinterpret data patterns for one reason or another. However, a study conducted last year by the University of Montreal confirmed that artificial intelligence algorithms can not only work as well as doctors, but actually make fewer errors in diagnosis.

This doesn’t mean that a doctor shouldn’t be present. In fact, the ideal way of approaching diagnosis with AI is to let the machine make an initial, educated guess and then have a human doctor confirm the diagnosis. This is good practice all around: it lets the computer learn in a controlled environment, and it makes the patient more comfortable than a fully automatic system would.

That being said, the fact that artificial intelligence algorithms are equally (or more) accurate at diagnosing patients does mean that in urgent cases where a doctor is unavailable, machine learning can provide a safety net of sorts.