As artificial intelligence advances, researchers are discovering new ways it can save lives, including eliminating harmful drug interactions, helping doctors diagnose rare forms of cancer, and tracking down criminals involved in human-trafficking rings.
At the University of California Medical Center in San Francisco, AI streamlines the process of filling prescriptions. Doctors and pharmacists input medication orders in a computer, and a machine picks, packages, and dispenses individual doses of pills, according to an article on the university’s website. Since 2010 the hospital has been using PillPick, a robotic system manufactured by Swisslog.
One important advantage to automating prescriptions is that computers are more likely than pharmacists to spot potential mistakes or drug interactions that could cause fatalities. A yearlong study of pharmacy errors published in 2015 by the American Journal of Health-System Pharmacy found that, among 50 hospital pharmacists, nearly five mistakes were made per 100,000 orders. UCSF reported that its PillPick, in its first few months, filled 350,000 prescriptions with no errors.
Researchers at the Houston Methodist Research Institute developed AI software last year that can accurately diagnose a patient’s breast cancer 30 times faster than doctors could.
When fed mammogram results and medical histories of about 500 patients, the software diagnosed breast cancer with 99 percent accuracy. And the software produces fewer false positives than doctors do: Currently about 20 percent of biopsies in the United States turn out to be unnecessary, according to the researchers.
AI has been found to diagnose other types of cancer as well, even when doctors haven’t. Last year the University of Tokyo’s Institute of Medical Science reported that IBM’s Watson supercomputer found a rare form of leukemia in a patient. The patient had been diagnosed with acute myeloid leukemia, and doctors were stumped when her post-chemotherapy treatment failed to work. The computer generated the correct diagnosis after comparing the patient’s genetic information with tens of millions of clinical oncology studies, according to a Japan Times article. The doctors changed the patient’s therapy, and her condition improved significantly.
HALTING HUMAN TRAFFICKING
More than 20 million people each year are sold worldwide into prostitution. Researchers at the University of California, Berkeley, have developed AI tools to identify sex-trafficking rings, making the leaders easier to target and prosecute, according to a Fortune article.
One way for law enforcement to combat sex trafficking is to parse vast numbers of online advertisements for sex, including classified-ad websites such as Backpage.
Berkeley’s team created a machine-learning filter that finds similarities among human-trafficking ads created by criminal organizations, making it easier to distinguish them from ads posted by prostitutes who are voluntarily engaging in the trade. And because Bitcoin is the most common payment option used by sex traffickers, the researchers developed a separate AI program that searches publicly available information on the cryptocurrency’s blockchain system. The program helps determine whether the ads were purchased by the same Bitcoin user, indicating they might link back to a human-trafficking ring.
The AI program found that one source had posted 10,000 ads in one month. And one Bitcoin wallet it investigated was used to pay for US $150,000 worth of sex ads, potentially revealing ties to a massive trafficking ring.