Avoiding bias in artificial intelligence
David A. Gudis, Edward D. McCoul, Michael J. Marino, Zara M. Patel
Abstract
Artificial intelligence (AI) is ubiquitous and expanding, and the healthcare industry has rapidly adopted AI and machine learning for numerous applications. It is essential to understand that AI is not immune to the biases that impact our clinical and academic work, and in fact may inadvertently amplify rather than reduce them. As we harness the power of AI, it is our obligation to our patients to ensure that we address these concerns. We must take responsibility for proactive stewardship to protect against bias, not only for new AI algorithms, but also for our research studies that may one day provide data for those algorithms.
Topics & Concepts
Stewardship (theology)MedicineObligationArtificial intelligenceApplications of artificial intelligenceMachine learningData scienceRisk analysis (engineering)Computer sciencePolitical sciencePoliticsLawArtificial Intelligence in Healthcare and EducationEthics in Clinical ResearchMachine Learning in Healthcare