In addition to improving the quality of care that health professionals can provide to their patients, of New England Journal of Medicine argues that artificial intelligence (AI) and machine learning will enable medical professionals to spend more time on essential human-to-human interactions.
In the 1990s and early 2000s, computer researchers were rapidly improving the ability of machines to perform repetitive medical tasks prone to human error. In recent decades, computer reading of electrocardiograms and differential leukocytes, analysis of retinal photographs and skin lesions, and other imaging tasks have become a reality, widely accepted, and incorporated into clinical practice.
“Scientists are using AI and machine learning to build massive grids of connected data to unlock new discoveries,” the authors write. “…these advances have enabled the emergence of computers that help us perform previously tedious tasks.”
But the use of AI and machine learning extends far beyond reading medical images to help identify outbreaks of infectious diseases. It combines clinical, genetic, and many other test results to identify conditions that may otherwise go undetected. Streamline the business operations of the health system. However, AI and machine learning still have important unsolved problems that need to be addressed before they can be used more safely and effectively in medicine.
First, the authors pointed out that the norms for using AI and machine learning have yet to be established. Researchers are still unsure about how the biases of the algorithms used to “teach” AI tools will affect them when applied to the real world. In many ways, human values seem to be superimposed on AI and machine learning tools, giving rise to the same problems faced by medical professionals rather than objective data.
Second, the exact role of AI and machine learning remains unclear, with countless potential uses proposed but yet to be implemented. For example, there are options to use AI and machine learning tools as personal notes or to use these tools to prompt doctors to ask key questions that may lead to differential diagnosis.
With all these options and countless others, more clinical studies are needed.The medical community expects the same amount of data and clarity with respect to AI and machine learning interventions as pharmaceutical interventions. However, the description and testing criteria for these tools are still unknown.
The authors say that research on AI and machine learning for medical applications requires three elements. Second, interventions must be definable, extensible and applicable. And finally, when the results are applied in practice, they should be beneficial to all patients under consideration, not just those who resemble those for which the algorithm was trained.
One particularly interesting example of the potential opportunities for AI and machine learning in healthcare is the rapid increase in the use of chatbots. The authors defined a chatbot as a computer program that uses AI and natural language processing to understand questions and create automated responses to them, simulating human conversation. A very early medical chatbot called ELIZA was developed between his 1964 and 1966, and recently chatbot technology has spread to almost every aspect of life.
“While chatbots have only just been introduced at a level of sophistication that could impact day-to-day medical practice, we believe they have considerable potential to impact the way healthcare is practiced,” the authors wrote. I’m here.
The new generation of chatbots are incredibly powerful and can be used as scribes and coaches for medical professionals. However, the authors point out some important caveats. In particular, chatbots can answer important questions that greatly help medical professionals, but it is difficult to know if the answers provided are based on proper facts, and it is difficult to calibrate and verify the work of chatbots. It is the clinician’s responsibility to
“Nevertheless, we believe chatbots will become an important tool in medical practice,” the authors write. “Like any great tool, they help us do our jobs better, but if not used properly, they can do harm.”
The authors conclude that the adoption of AI and machine learning has definitely helped healthcare professionals improve the quality of care they can provide to their patients, and has incredible potential to further improve care. attached. Rather than putting medical professionals out of business, as the authors are concerned, linking them with these tools will help them do their jobs more efficiently and make important interactions with patients. can be done, which makes the medical profession very rewarding.
Haug CJ and Drazen JM. Artificial Intelligence and Machine Learning in Clinical Medicine, 2023. New Engl J Med2023;388:1201-1208. doi:10.1056/NEJMra2302038.