The use of artificial intelligence (AI) in our lives, both in peace and in war, has exploded over the past 10 years, from ATM machines armed with facial recognition to Alexa, your friendly companion which can answer questions, let you know the weather for the day and turn on your television. The problem is that Alexa, among other machines, stores and reports your questions and formulates a profile of you and your interests. AI has been steadily saturating our lives. But as always with a novel technology, predictions for its power and how it can be used are often more fantastical than real. AI’s limitations and possibilities for error are rapidly becoming apparent. While speculation is fascinating, constructing, developing and using AI technology and testing its accuracy in replacing human effort requires careful, evidence based research.
Let’s talk about using AI in the doctor’s office. Patients complain bitterly about the physician who looks into his computer screen and types furiously to create the latest buzz word in our profession, the “electronic medical record. ” The record is stored and, like any digital information, can be hacked and utilized for nefarious purposes. Often it contains wrong data that are copied verbatim and reported time after time as the records are transmitted from one physician to the next. Another iteration of the medical record is the product of the “medical scribe”, a robot that records what is said between physician and patient. Unfortunately, the scribe may “hear” incorrectly (ever follow the instant caption service on TV and see how often it misses the original spoken word?) and often cannot recognize regional accents.
Take facial recognition technology: sounds like a fail-safe idea for secure access to cell phones, ATM’s and at airports. But it seems that the current systems can’t take skin tone into account-a source of misidentification.
Alexa has partnered with the FDA to recommend and keep track of medications for patients but the machine doesn’t recognize many of the names of the drugs ordered/taken.
Robots are everywhere in the world of commerce. The New York Times reports that a company called Powerfront “uses colorful avatars to represent customers, their purchasing history and their shopping activity.” Business knows more about you than you can imagine, tracking your internet use and storing the information for marketing purposes. One expert believes customer’s “emotional profiles” are stored and used to sell products they may find particularly attractive and even irresistible.
The Washington Post reported this week that the Massachusetts State Police has become the first law enforcement agency in the US to put a robotic dog to work. They are testing the capabilities of the machine, which they have attached to be bomb squad unit, for three months to detect “potentially dangerous environments which may contain suspects and ordnance”. Civil rights organizations are becoming alarmed. Apparently much of the robot’s movements are self-generated and not controlled by humans. The possibility of physical harm or intimidation of people has apparently been a concern and been expressly forbidden by agreements made by the robot manufacturer, Boston Dynamics, and their customers.
The bottom line is that artificial intelligence is accumulating information about all of us to a degree that’s intrusive, alarming and can contain significant errors. It is being used to influence choices and purchases we make to an extent that most of us would find astonishing if we had an opportunity to review our individualized profiles. Machines aren’t ready to replace the judgment and control of human intelligence without close supervision and clear boundaries of what’s transparent, fair and legal. Before you consent to “cookies” the next time you access a new website, think twice.
Dr. Marianne Legato, Professor Emerita of Clinical Medicine at Columbia University is an internationally known academic physician, author, lecturer, and specialist in gender-specific medicine. She is founding member of the International Society for Gender Medicine and also the founder and director of The Partnership for Gender-Specific Medicine at Columbia University and its next iteration, The Foundation for Gender-Specific Medicine. These enterprises are the first collaborations between academic medicine and the private sector focused solely on gender-specific medicine: the science of how normal human biology differs between men and women and of how the diagnosis and treatment of disease differs as a function of gender and sex. Her ground breaking textbook on Gender-and Sex Specific Medicine has been published in 2017 in the 3rd edition.
She has published extensively on Gender and Sex Specific Medicine, both scientifically and for the lay public. She is also the founding editor of the journal Gender Medicine, and the Journal Gender and Genome, published for the scientific community. In 1992, Dr. Legato won the American Heart Association’s Blakeslee Award for the best book written for the lay public on cardiovascular disease. She is a practicing internist in New York City and has been listed each year in New York Magazine’s “Best Doctors” since the feature’s inception in 1993.