Evaluating the Reliability of Artificial Intelligence in Healthcare: The Doctors’ Perspective in Northern Greece

Machine learning (ML) refers to the ability of machines to advance their performance by relying on previous results or observations. ML algorithms empower computers to learn without the need to be programmed. In recent years ML algorithms, as a branch of Artificial Intelligence (AI), have been employed in many fields and are used in multiple applications towards improving daily human lives through efficient data mining, and thus, trustful decision-making. This work attempts to examine the reliability of ML, referring to AI in the broader sense, in healthcare, currently offering techniques and methods that contribute to the clinical diagnosis of patients, as perceived by doctors. The study explores the concepts of trustworthiness of clinical AI systems in healthcare, followed by a survey on experts’ opinions in the health sector regarding their trust and the factors that mainly affect it. The survey includes a target group of 35 doctors located in Northern Greece and a corresponding questionnaire. Research identified nine factors that influence the medical staff towards the overall adoption of AI in healthcare. Moreover, the drawn conclusions based on the analysis of the survey’s results revealed that doctors in Northern Greece consider AI systems as essential supporting medical tools and are positive about using them, even though some find them costly, complex, and still don’t fully trust them.