Bioengineer.org11/19 12:45 UTC
Narrative Review Identifies Systematic Bias and Privacy Risks Across AI Electrocardiography Protocols
A narrative review of 46 studies finds that ethical vulnerabilities in artificial intelligence electrocardiography stem from upstream study protocols rather than algorithms alone.
Read the original →Bioengineer.org11/19 14:00 UTC
Granular Skin Tone Scales Expose Hidden Bias in AI Bruise Detection
A study shows that granular skin tone scales reveal AI bruise detection disparities. Broader classification methods conceal these disparities.
Read the original →Bioengineer.org11/19 11:10 UTC
Moral Diversity in Clinical Care Complicates Ethical Responsibility for Medical Artificial Intelligence
A conceptual analysis finds that diverse moral frameworks among clinicians and patients disrupt the traditional attribution of responsibility in AI-guided clinical decision support systems.
Read the original →Bioengineer.org11/19 02:11 UTC
Review Finds Fraud Detection AI Prioritizes Accuracy Over Fairness and Explainability
A systematic review of AI models for credit card fraud detection shows high benchmark accuracy. It also finds widespread deficits in transparency, imbalance handling, and demographic fairness auditing.
Read the original →