Most patients believe they should about whether AI is used for reporting their images, according to survey results published June 30 in Radiology.
And for incorrect imaging results, many patients believe that both AI and radiologists are at fault, wrote a team led by Hayley Briody from Beaumont Hospital in Dublin, Ireland.
“Overall…transparency appears important when implementing AI tools in radiology to ensure patient comfort with their use,” the Briody team wrote.
While attitudes about AI from radiologists and clinicians have been documented, Brody and colleagues noted how patient opinions about the technology are “under-evaluated.” The researchers at RSNA 2025 presented results suggesting that most patients support AI’s use in this area, but not as a standalone reader.
Briody and colleagues studied patient attitudes about AI use being disclosed and accountability in radiology. They used the same data from the previous study, which included survey data from 1,041 patients. Most respondents were aged 50 to 69 years (42.4% and of white Irish ethnicity (85.4%).
The team reported the following:
96.0% of respondents (993/1,034) said patients should be informed about the use of AI for reporting images.
Over half of respondents (53.4%) prefer written communication while about one-third of respondents (33.6%) prefer verbal consent.
Nearly two-thirds of respondents (656/1,024) said both the radiologist and AI tool are at fault for incorrect results provided after reading by both.
The results reflect a preference for transparency by patients, the study authors highlighted. They also noted the Blueprint for an AI Bill of Rights document by the White House Office of Science and Technology Policy in the U.S., which stresses a “right to notice and explanation.” However, this does not clearly define whether this communication should take the form of standardized notification or informed consent.
“We hope this work prompts further research to determine whether attitudes differ when AI tools are used for secondary purposes, such as triaging radiology requests,” they wrote.
Read the full study here.
