Researchers developed an AI model that detects transthyretin amyloid cardiomyopathy (ATTR-CM) directly from standard 12-lead ECG images, achieving 87% accuracy (AUROC 0.87) across eight multinational validation cohorts. The algorithm maintained consistent performance across diverse populations in the US and Europe, including high-risk groups like older Black and Hispanic adults with heart failure. ATTR-CM is a treatable but severely underdiagnosed cause of heart failure, often detected only after irreversible cardiac damage occurs. This represents a potential breakthrough for early detection, as ECGs are universally available and frequently performed years before traditional diagnosis. The technology could democratize screening access, particularly benefiting underserved populations who lack access to specialized cardiac imaging and expertise. Early identification enables timely initiation of disease-modifying therapies that can slow progression and improve outcomes. However, this remains a preprint awaiting peer review, and real-world implementation would require validation of clinical workflows, cost-effectiveness analysis, and integration with existing healthcare systems. The approach exemplifies how AI can transform routine diagnostic tools into powerful screening instruments for rare but serious conditions.