Knowledge, attitudes and practices towards artificial intelligence (AI) among radiologists in Saudi Arabia

Efficiency, accuracy, and health professional's perspectives regarding artificial intelligence in radiology practice: A scoping review

“…In the remaining articles, AI was used in detecting thyroid nodules [36], acute fractures [37], and intracranial hemorrhage [38] as well as checking endotracheal tube (ETT) placement [39]. Thirty‐six studies assessed radiologists' attitudes or views regarding the application of AI in radiology, including the opinions of radiologists ( n = 26) [12–14, 40–62], general health care professionals ( n = 1) [63], residents ( n = 4) [64–67], and physicians ( n = 1) [35]. Several studies in our review examined the viewpoints of multiple health professional groups.…”

“…Additionally, four studies reported that most radiology specialists lacked education related to AI [42, 49, 52, 54]. Most participants (25.6%–98.7%, median = 84.8%) expressed approval with respect to integrating AI‐related content into the educational curricula of radiology residents [12, 44, 48, 49, 52, 59, 61, 63, 64, 68]. Many respondents wanted to keep learning about AI, particularly its clinical applications, pros and cons, and technical methods [14, 45, 48, 49, 52, 65, 66].…”

“…Some studies found that trust in AI was affected by the task or application, level of automation, perceived accuracy of the algorithm, and radiologists' experience and training [8]. Seventeen surveys specifically focusing on replacement of radiologists with AI suggested that a median 71.1% of participants did not believe that AI will significantly impact job security within radiology [12, 14, 41–44, 47, 52, 53, 58, 59, 61, 63–66, 68]. Interestingly, some residents indicated a preference for radiology as a future career owing to the potential of AI [52, 64, 65].…”

“…Six studies reported that 20.4%-72.0% of radiologists have used AI [45,58,63,65,67,68]. Most studies reported positive views among radiologists toward incorporating AI in clinical practice [8,52,55,57,59,60,64,[66][67][68]. Specifically, 29.5%-94.3% (median = 67.4%) of respondents in seven studies said they believed that AI has the potential to enhance clinical accuracy in radiology practice [12,14,40,41,46,63,66,67].…”

“…Many respondents (22.8%-85.0%, median = 70.6%), represented in 12 studies, had positive attitudes toward AI and viewed it as an essential component of future radiology practice [14,35,41,42,50,56,59,60,64,[66][67][68]. Five studies found that 11.7%-91.2% (median = 27.9%) of respondents expressed confidence in AI diagnostic results [35,43,45,61,63].…”