Professor on using AI to detect medical errors, the future of AI in medicine – The Stanford Daily

Scripps Professor of Molecular Medicine Eric Topol gave a talk on the future of artificial intelligence (AI) in medicine at the Stanford Artificial Intelligence in Medicine & Imaging Symposium on Aug. 5, where he discussed developing medicine that is much more accurate than it is now. According to Topol, there are over 12 million serious medical errors per year, leading to added costs and other complications.

In one sense, machines are much more capable than humans at detecting errors, Topol said. Machines are able to find different medical complications that doctors themselves may not be able to find, such as tumor nodules, and can advance medical technology, as with mammographies.

If you show this picture [of a retina] to retinal authorities, their chance of getting [the gender] right is 50%, Topol said. But interestingly, an AI algorithm can be trained to be 97 to 98% accurate.

Machines can identify many other aspects of a patient simply from the retina, including ones possibility of kidney disease, Alzheimers and diabetes and age and sex. Topol said applications of AI are not limited to images of the retina: AI can interpret all sorts of slides and images that pathologists cannot see or make conclusions.

This is going to affect, in the future, every medical discipline, he said. Theres no exception here. And its not just those in the hospital. Its paramedics, its all the paraprofessionals, pharmacists the works.

Topol said AI would transform the health and medical field with tasks such as interpreting scans, selecting embryos for in vitro fertilization and predicting death in hospitals. AI will also greatly improve the resolution and quality of medical scans.

Machines will not replace physicians, Topol said, quoting professor and neurosurgeon Antonio Di Leva. But physicians using AI will soon replace those not using it.

Even with the tremendous potential of AI applications, Topol still emphasized that humanistic medicine is important in the long-term advancement of AI in medicine.

Even now, Topol said that patients often feel rushed during their doctor visits. Its not uncommon to see physicians typing on computers or facing their screens, without making eye contact with patients during doctor visits.

Topol said in medicine there is a lack of empathy, which companies and institutions should focus on improving.

We want to restore the humanistic side the critical component of what is medicine, Topol said.

Contact Amanda Zhu at amandaz9888 at gmail.com.

Visit link:
Professor on using AI to detect medical errors, the future of AI in medicine - The Stanford Daily