Submitted:
30 September 2026
Posted:
04 October 2026
You are already at the latest version
Abstract
Background/Objectives: Healthcare professionals have shown different perceptions regarding the safety, validity, trust, and expectations associated with the implementation of medical AI in the world. This study aimed to assess the knowledge, attitudes, and practices concerning artificial intelligence (AI) among healthcare professionals at Jendouba Hospital, northwest Tunisia. Methods: This was a cross-sectional study realized between September and December 2025, targeting all medical healthcare workers in Jendouba Hospital. Data were collected through a well-structured questionnaire. Results: A total of 224 healthcare workers responded (response rate: 90.6%). Most participants were unable to differentiate between core AI concepts. Over half supported integrating AI into formal training programs, while 61.5% strongly disagreed that teachers could be replaced by AI. Additionally, 60.5% believed AI cannot function without human assistance, and 52.2% disagreed that clinical AI will outperform human professionals. Although more than half found AI often useful, over a quarter reported being misled by it at times. In research, AI was mainly used for grammar and spell-checking. A majority (88.3%) believed that AI should always be supervised by physicians. Concerning ethics, nearly half felt that AI could compromise the doctor–patient relationship, while 42.4% did not consider its use a breach of medical confidentiality. Mean attitude and practice scores were significantly higher among professionals in non-clinical specialties compared to clinical ones. Conclusion : While healthcare staff favor AI integration despite limited technical knowledge, ethical concerns highlight the need for targeted training. These findings offer vital baseline data to foster equitable AI implementation in developing healthcare systems.