3-minute article

How AI works in medicine

Accurately diagnosing disease may be a challenge in some circumstances, depending on the patients and multiple factors that may be at play. In fact, one study estimates that diagnostic errors affect at least one in 20 adults in the United States, according to BMJ Quality & Safety. But with the advances in technology doctors and hospitals are enlisting the help of machines combined with artificial intelligence (AI) in diagnosing illnesses.

AI works through a method called machine learning which uses algorithms that are “fed” specific data. As an algorithm analyzes the data to perform its task, it learns, becoming better at solving a task over time. A more advanced type of machine learning called deep learning, which has more capabilities to learn more complex tasks like processing images and sound, for example.*

Machine learning and deep learning both play a role in computer-aided diagnostics (CAD) to assist medical professionals. To better understand how CAD works, it’s helpful to look at one disease as an example — breast cancer. In the United States, it is estimated that more than a million surgical breast biopsies are performed, and only 15% – 30% of the samples are malignant, according to an article published in Medscape. That means a large number of women are subjected to unnecessary biopsy. Doctors (and insurance companies) understandably want to reduce the number of unnecessary biopsies and hope CAD can help.

Man + machine
In various studies where both doctors and CAD look at radiology tests (like CAT scans, mammograms and MRIs), a partnership between both man and machine yields the best results.

According to a large international study from Google, Northwestern Medicine and two screening centers in the United Kingdom (U.K.), a new AI model is shown to predict breast cancer in mammography scans more accurately than radiologists, reducing false positives and false negatives. With the assistance of CAD and this deep-learning AI model, doctors are able to better assess whether a tumor might be malignant. In the study, the AI model resulted in a 9.4% reduction in false negatives and a 5.7% reduction in false positives among the U.S. study participants.

Ultimately, studies like this one are showing that the best results come from a partnership between AI (initially programmed by doctors) and doctors themselves. “In some examples, the human outperforms the AI and in others, it’s the opposite. But the ultimate goal will be to find the best way to combine the two …,” states Northwestern study co-author Dr. Mozziyar Etemadi.

Broader applications
Not every scenario allows people and machine to work together. In some remote regions of the world, for instance, where there’s a shortage of doctors, AI may be able to make great strides in detecting treatable diseases like tuberculosis.

The balance between man and machine also has applications outside of medicine. For instance in data science, machines can process infinitely more information than humans, but AI may also miss subtleties that humans can pick up on, like context or tonality.

For instance, algorithms programmed to react to tweets or news headlines may not detect sarcasm or jokes and could react incorrectly. Similarly, driverless cars may help eliminate human error in some situations, but humans may also be able to detect dangers that the car’s sensors cannot.

Working together to help clients reach their financial goals
People and machines using AI technology are also an important team in finance. Many algorithms working in fields like finance are constantly being monitored and evaluated by humans. While machines “learn” on their own, continued human monitoring helps ensure they’re getting the right inputs. This combination can help clients achieve their financial goals for retirement by optimizing potential outcomes.

Two things you can do today:

  1. Think about how machines and AI might be able to help you improve the outcomes at work.
  2. Discover more about Athene’s new indices that optimize technology to help improve financial outcomes for customers.

This information is brought to you by Athene — where innovative annuity solutions are powered by unconventional thinking.