Artificial intelligence (AI) has seen exponential advancements, rapidly approaching the development of Artificial General Intelligence (AGI) and Artificial Super Intelligence (ASI). As these technologies mature, ethical challenges around bias, fairness, and moral accountability become increasingly critical. Current governance frameworks, such as the NIST AI Risk Management Framework and the EU AI Act, offer important regulatory steps but have limitations when addressing the nuanced ethical issues associated with AGI and ASI. To bridge these gaps, this study introduces Virtue Ethics by Design (VED), an innovative approach advocating for the integration of virtue ethics directly into AI architectures. VED proposes embedding virtues, such as fairness, honesty, and prudence within AI systems through specialized data curation, virtue-oriented algorithmic design, and adaptive ethical testing frameworks. Drawing from Aristotle’s virtue ethics and practical implementations like IBM’s AI Fairness 360 toolkit, this approach ensures AI not only performs efficiently but also inherently aligns with human values. Implementing VED could significantly enhance the ethical robustness of future AGI and ASI systems, contributing positively to societal welfare and promoting sustainable human-AI coexistence.