While patients are excited about the potential of healthcare artificial intelligence in theory, when it comes down to the real thing most have major concerns, according to a May 4 study published in JAMA Network Open.
More than 900 patients from the New Haven, Conn.-based Yale Cancer Center took part in the study in December 2019, sharing their worries, hopes and opinions on healthcare AI.
The majority of patients had faith in health AI in medicine, with 55.4 percent saying it would make healthcare much or somewhat better and only 6.2 percent said it would make healthcare much or somewhat worse.
However, in practice, the black box problem of AI made patients uneasy, with over 70 percent of patients saying receiving a diagnosis from an AI without a clear rationale made them uncomfortable in some way. Concerns were brought up about a range of other factors too, with 91.5 percent saying they were concerned about misdiagnosis with AI machines, 70.8 percent were worried about privacy breaches and 69.6 percent thought they'd get less time with physicians.
"In many ways, our work highlights a potential blind spot among AI researchers, which needs to be addressed as these technologies become more common in clinical practice," said Sanjay Aneja, MD, a senior author of the study. "Patient education, concerns, and comfort levels should be taken into consideration when planning for integration of AI."