Participants’ Perceptions of Privacy and Data Sharing Regarding Health-Related Data Using Artificial Intelligence

Abstract

Substantial research has been performed to explore the effective integration of artificial intelligence (AI) into healthcare. Consequently, it becomes imperative to grasp the perspectives and concerns of general public in relation to the utilization of their health data in AI research, particularly within the context of privacy issues. This study aims to explore the awareness of participants about the AI, their privacy concerns about data sharing using AI and its use on healthcare data. We carried out a comprehensive study by employing a self-administered questionnaire tool with participants using convenience sampling methodology. A total of 450 participants were enlisted from Saudi Arabia. Conditional binary logistic regression models were employed to compute odds ratios (ORs) and 95% confidence intervals. Among the participants, 168 (37.3%) showed that they have knowledge about AI. In terms of personal data’s vulnerability when using AI technology, 186(41.3%) perceived a privacy risk about their health data. 201(44.7%) indicated their trust in AI’s ability to safeguard data privacy. Regarding the use of machine learning for medical record analysis, 180(40.0%) declared it riskier than benefits. For AI research purposes, 205 (45.6%) supported data sharing and 213(47.3%) believed hospitals should have strict regulations while 214 (47.6%) believed hospitals should provide limited access to data to ensure health data privacy. Furthermore, the study found that younger individuals were more likely to trust AI with their data privacy (OR = 0.540, 95% CI; 0.300-0.972), while participants with higher education levels were nearly three times more likely to trust AI with their data privacy compared to those with lower education (OR = 2.894, p = 0.047, 95% CI; 1.012-8.278). Patients’ viewpoints, the extent of assistance they receive, and their comprehension of health data research and artificial intelligence exhibited significant variations, often influenced by privacy issues. To ensure the acceptability of AI research and its seamless integration into clinical practice in the future, it is imperative to engage the public more extensively and stimulate discussions, particularly concerning privacy concerns.